A flexible traction substation energy management method for photovoltaic and energy storage system access

Abstract

The application discloses a kind of photovoltaic and energy storage system access's flexible traction substation energy management method, comprising the following steps: the energy management scheme is divided into day, day and real-time three time scales;In day optimization scheduling stage, based on train traction load and photovoltaic output short-term prediction data, with the lowest traction substation comprehensive operation cost as objective function, and with guaranteeing traction substation normal operation constraint condition is built, the mixed integer linear programming model of day flow optimization scheduling is established;In day rolling optimization stage, based on model predictive control method, with the minimum traction substation power supply plan deviation as objective function, and with control time domain substation normal operation constraint condition is built, the rolling correction of energy storage system work plan is completed;In real-time optimization operation stage, consider photovoltaic flexible output scheme, and establish the nonlinear optimization model of the lowest traction substation original side negative sequence current, solve the real-time power compensation instruction of back-to-back converter.

Description

Technical Field

[0001] This invention belongs to the field of energy management technology for traction power supply systems, and specifically relates to an energy management method for flexible traction substations with photovoltaic and energy storage systems integrated. Background Technology

[0002] With the successive introduction of development plans for modern urban financial districts across the country, passenger travel demand has also risen rapidly. Consequently, the scale of electrified railway construction in my country has continued to grow, with a planned long-term railway network of 200,000 kilometers, including 70,000 kilometers of high-speed rail. However, while electrified railways are developing rapidly, their energy consumption is also increasing. According to the National Railway Administration's statistical bulletin, in 2021, the total electricity consumption of the national railway reached 78.7 billion kilowatt-hours, equivalent to 15.807 million tons of standard coal, a year-on-year increase of 5.74%. This has led to a year-on-year increase in railway operating costs. At the same time, the enormous traction power of high-speed railways (for example, the CRH380AL EMU has reached 20MW) has exacerbated the negative sequence current problem, hindering the smooth interaction between traction substations and the three-phase power grid.

[0003] Against this backdrop, integrating photovoltaic (PV) power generation and energy storage systems into traction substations to effectively utilize new energy power generation resources along railway lines and regenerative braking energy of trains, thereby reducing the overall operating costs of traction substations and compensating for negative-sequence current, has become a research hotspot in the railway industry. However, the drastic fluctuations in train traction load (reaching 10MW within seconds) and the uncertainty of PV output pose significant challenges to the system's energy management solutions. Therefore, for traction substations integrating PV and energy storage, there is an urgent need to develop energy management solutions that can overcome the randomness of traction load and PV output, achieve economically optimized operation of the traction substation, possess real-time negative-sequence current compensation capabilities, and enable friendly interaction with the three-phase power grid. Summary of the Invention

[0004] In order to solve the technical problems existing in the background art, the present invention aims to provide an energy management method for flexible traction substations with photovoltaic and energy storage systems connected. This method can overcome the adverse effects of prediction errors in traction load and photovoltaic output, improve the economic optimization operation capability of traction substations, and effectively realize real-time compensation of the primary side negative sequence current of the substation, which is conducive to friendly interaction with the three-phase power grid.

[0005] To solve the technical problem, the technical solution of the present invention is as follows:

[0006] An energy management method for flexible traction substations with integrated photovoltaic and energy storage systems divides the system energy management strategy into three time scales. The method includes:

[0007] S1: In the day-ahead optimization scheduling phase, based on the short-term forecast data of train traction load and photovoltaic output, the objective function is to minimize the comprehensive operating cost of the traction substation, and the constraint condition is to ensure the normal operation of the traction substation. A mixed integer linear programming model for day-ahead power flow optimization scheduling is established.

[0008] S2: In the intraday rolling optimization stage, the objective function is to minimize the deviation of the power supply plan of the traction substation, and the constraint conditions are constructed by controlling the normal operation of the substation in the control time domain to build a model predictive control scheme for the flexible traction substation.

[0009] S3: In the real-time optimization operation phase, consider the photovoltaic flexible output scheme and establish a nonlinear programming model with the lowest primary-side negative sequence current of the traction substation. Solve the real-time power compensation command of the back-to-back converter and construct a real-time optimization operation method for the flexible traction substation.

[0010] Furthermore, prior to step S1, the method further includes: dividing the energy management method into three time scales: day-ahead, intraday, and real-time.

[0011] Furthermore, by employing the traction load and photovoltaic forecasting method, short-term forecast data of train load and photovoltaic output on both sides of the traction substation during the dispatching day are obtained.

[0012] The objective function for establishing the mixed-integer linear programming model for day-ahead power flow optimization scheduling is as follows:

[0013]

[0014] In the formula: To schedule the total daily operating cost; , These are the electricity purchase price for traction substations and the electricity cost for returning three-phase power to the grid, respectively. The price is based on demand. , These are the operation and maintenance costs of energy storage and photovoltaic systems, respectively. , These represent the power purchased from the three-phase power grid and the power transmitted back to the traction substation at time t, respectively. This represents the maximum required power. , These represent the charging and discharging power of the energy storage system at time t, respectively. Let t be the power output of the photovoltaic power generation system at time t; T is the number of scheduling periods in the day-ahead power flow optimization scheduling phase. This refers to the time scale used in the current optimized scheduling phase.

[0015] Furthermore, based on the flexible traction substation model and short-term forecast data of train load and photovoltaic power output, constraints for the day-ahead power flow optimization scheduling stage are established, including: substation power balance constraints, primary-side common coupling point (PCC) power constraints, back-to-back converter capacity constraints, energy storage system charging and discharging power constraints, energy storage medium state of charge constraints, and maximum demand power constraints.

[0016] Furthermore, after establishing a mixed-integer linear programming model for the day-ahead power flow optimization scheduling stage, the model is solved to obtain the substation PCC point power plan, the energy storage system charging and discharging power plan, and the energy storage medium state of charge plan. This completes the day-ahead power flow optimization scheduling of flexible traction substations with photovoltaic and energy storage system access, and the above scheduling plans are issued in advance to the intraday rolling optimization operation stage.

[0017] Furthermore, during the intraday rolling optimization operation phase, a model predictive control method for time k of the flexible traction substation is established, specifically including:

[0018] A vector is formed by the PCC point power, energy storage system charging and discharging power, and state of charge at time k in the flexible traction substation. The state vector is selected; the increment of the energy storage system output at time k in the flexible traction substation relative to the planned value before the day is chosen. The control vector is formed by selecting the increments of the ultra-short-term predicted power curves of the total train load and photovoltaic output of the left and right power supply arms at time k of the flexible traction substation relative to the day-ahead predicted curve. The disturbance vector is selected; the vector is formed by the power at point PCC of the flexible traction substation at time k and the state of charge of the energy storage system. As the output vector, a prediction model for the predictive control method of the flexible traction substation at time k is established:

[0019]

[0020] In the formula: For the charging and discharging efficiency of the energy storage system; This is the time scale for the intraday rolling optimization phase.

[0021] Furthermore, ultra-short-term forecast data of traction load and photovoltaic output are obtained from time k+1 to time k+P, i.e., within the prediction time domain. The prediction model described above iterates step by step within the ultra-short-term prediction time domain to construct the output vector matrix in the prediction time domain as follows:

[0022]

[0023] In the formula: (p=1, P) is the output vector of the prediction model at time k, which is composed of the power of the substation PCC point and the state of charge of the energy storage system at time k+p.

[0024] Based on the acquired ultra-short-term forecast data of traction load and photovoltaic output, the control time domain is selected from time k+1 to time k+M of the flexible traction substation, where the control time domain must be less than or equal to the prediction time domain, i.e., M≤P. The optimization objective of the model predictive control method at time k is as follows:

[0025]

[0026] In the formula: The tracking plan target matrix for controlling the output vector in the time domain at time k is given by the power values ​​of the PCC points in the daily plan at each time point in the time domain. (p=1, M), energy storage state of charge (p=1, Composed of (M); The control matrix at time k is the control matrix within the time domain, which is the increment of energy storage power relative to the day-ahead plan at each time point in the time domain. (p=1, The matrix consists of M and W, which are weight coefficient matrices.

[0027] Furthermore, based on the flexible traction substation model and the ultra-short-term forecast data of load and photovoltaic output obtained above, constraints are established for the flexible traction substation under normal operating conditions in the control time domain, including: substation power balance constraints, primary side common coupling point power constraints, back-to-back converter power constraints, energy storage system charging and discharging power constraints, energy storage system state of charge constraints, and maximum demand power constraints.

[0028] Based on the objective function of the prediction model of the above-mentioned model predictive control method and the normal operation constraints of the flexible traction substation in the control time domain, a quadratic programming model for the intraday optimized operation stage at time k is established. The quadratic programming model is solved to obtain the charging and discharging power correction vector of the energy storage system at times k+1, k+2, ..., k+M. The correction amount in the first time period of the vector, i.e., from time k+1 to k+2, is then sent to the real-time operation control system of the energy storage system.

[0029] Furthermore, in the real-time optimization operation phase, a real-time optimization operation method for the flexible traction substation at time k is constructed, specifically including:

[0030] The correction amount of the energy storage system from k to k+1 time obtained by the model predictive control method at k-1 time of flexible traction substation is superimposed with the day-ahead work plan value of the energy storage system from k to k+1 time to obtain the final work plan of the energy storage system from k to k+1 time.

[0031] Real-time acquisition of the voltage of the left and right power supply arms during time k to k+1. and Load current of the left and right power supply arms and and the output voltage and current values ​​of the photovoltaic array and Calculate the active power of the total train load on both the left and right power supply arms. and reactive power and photovoltaic power generation It then determines whether there is any power feedback from the PCC point to the three-phase grid at the current moment. If there is no feedback, the photovoltaic grid-connected converter operates in maximum power point tracking mode; otherwise, it adopts power-limiting output mode. The real-time output power of the photovoltaic is as follows:

[0032]

[0033] In the formula: The final working plan value of the energy storage system obtained above.

[0034] Based on the real-time acquired loads of the left and right power supply arms and the photovoltaic output power obtained after control by the photovoltaic grid-connected converter, the objective function for real-time optimization compensation of negative sequence current is established as follows:

[0035] In the formula: k T For the traction transformer turns ratio; , These represent the total active power of the α and β phase power supply arms at time n, respectively. , These represent the total reactive power of the α and β phase power supply arms at time n, respectively.

[0036] Based on the flexible traction substation model and the obtained real-time load data and photovoltaic control output data, constraints are established for the back-to-back converter under normal operating conditions, including: power balance constraints of the left and right power supply arms, power balance constraints of the DC bus, and capacity constraints of the back-to-back converter.

[0037] Based on the objective function for real-time optimization compensation of negative sequence current and the normal operation constraints of the back-to-back converter obtained above, a nonlinear programming model for the intraday optimization operation phase at time n is established. This model is then solved quickly to obtain the real-time active power compensation command of the back-to-back converter at time n. and Real-time reactive power compensation command and The instruction is then sent to the real-time operation control system of the back-to-back converter.

[0038] Furthermore, the method also includes:

[0039] In the ultra-short-term forecast data of traction load and photovoltaic output from k+2 to k+P+1 time, a model predictive control scheme for flexible traction substation at k+1 time is established. The model predictive control method for flexible traction substation at k time is repeatedly established in the intraday rolling optimization operation stage. A quadratic programming model for the intraday optimization operation stage from k+2 to k+M+1 time is established. The model is solved to obtain the charging and discharging power correction vector of energy storage system at k+2, k+3, ..., k+M+1 time. The correction amount in the first time period of the vector, i.e., from k+2 to k+3 time, is sent to the real-time operation control system of energy storage system. This process is continuously rolled forward for optimization.

[0040] The correction amount of the energy storage system from time k+1 to time k+2 is obtained from the model predictive control scheme of the flexible traction substation at time k, and superimposed with the day-ahead working plan value of the energy storage system from time k+1 to time k+2 to obtain the final working plan of the energy storage system from time k+1 to time k+2. The real-time optimization operation method of the flexible traction substation at time k+1 is constructed by repeating the real-time optimization operation stage, and the real-time optimization operation plan from time k+1 to time k+2 is solved to obtain the real-time output power command of photovoltaic power, the real-time active power and reactive power compensation command of back-to-back converters from time k+1 to time k+2, and then sent to the operation control system of each converter.

[0041] Therefore, as k increases, the process is repeated to complete the energy management method for flexible traction substations that integrate photovoltaic and energy storage systems.

[0042] Compared with the prior art, the advantages of the present invention are as follows:

[0043] 1. This invention divides the energy management method of flexible traction substations with photovoltaic and energy storage systems into three time scales: day-ahead, intraday, and real-time. Day-ahead power flow optimization scheduling realizes the effective utilization of photovoltaic power generation resources and train regenerative braking energy by traction substations, and improves the economic operation capability of traction substations under the two-part toll scheme.

[0044] 2. The intraday rolling optimization stage of this invention considers the impact of the uncertainty of traction load and photovoltaic output on the operation plan, reduces the deviation of operation plan caused by prediction data error, and improves the energy consumption and maximum demand reduction capabilities of flexible traction substations.

[0045] 3. The real-time optimized operation phase of this invention considers the real-time optimization compensation method for negative sequence current under rapid fluctuations in traction load, which reduces the negative sequence current on the primary side of the substation and meets the real-time operation control requirements of the traction substation. Attached Figure Description

[0046] Figure 1 This is a structural diagram of the flexible traction substation where photovoltaic and energy storage systems are integrated in this invention;

[0047] Figure 2 This forms the basic framework of the multi-timescale energy management scheme in this invention;

[0048] Figure 3 This is a flowchart of the photovoltaic flexible output scheme during the real-time optimized operation phase of this invention;

[0049] Figure 4 This is a comparison chart of the PCC point power and its reference planned power in the day-ahead power flow optimization scheduling method in the embodiment.

[0050] Figure 5 This is a comparison diagram of the PCC point power of the present invention and its reference planned power in the embodiment. Detailed Implementation

[0051] The specific implementation of the present invention is described below with reference to embodiments:

[0052] It should be noted that the structures, proportions, sizes, etc. shown in this specification are only used to complement the content disclosed in the specification for those skilled in the art to understand and read, and are not intended to limit the conditions under which the present invention can be implemented. Any modifications to the structure, changes in the proportions, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0053] Furthermore, the terms such as "upper," "lower," "left," "right," "middle," and "one" used in this specification are merely for clarity of description and are not intended to limit the scope of the invention. Any changes or adjustments to their relative relationships, without substantially altering the technical content, should also be considered within the scope of the invention.

[0054] Example 1:

[0055] The flexible traction substation structure targeted by this invention is as follows: Figure 1 As shown, the basic framework of the energy management scheme for flexible traction substations with photovoltaic and energy storage systems is as follows: Figure 2 As shown.

[0056] The present invention specifically includes the following steps:

[0057] Step 1: During the day-ahead power flow optimization scheduling phase, a recurrent neural network prediction model is used to obtain the predicted active power of the train load on the left and right sides of the traction substation during the scheduling day; meteorological data and solar radiation data for the scheduling day are obtained from the local meteorological department, and a direct prediction model is used to obtain the predicted photovoltaic output data.

[0058] Step 2: Considering the two-part electricity pricing method for electrified railways, the electricity pricing parameters for traction substations include electricity cost, return electricity cost, and demand cost. The operation and maintenance costs for flexible traction substations with photovoltaic and energy storage systems include energy storage system operation and maintenance costs and photovoltaic power generation system operation and maintenance costs. Based on the electricity cost and operation and maintenance cost collection scheme for traction substations, the objective function for the day-ahead power flow optimization scheduling scheme is established as follows:

[0059]

[0060] In the formula: To schedule the total daily operating cost; , These are the electricity purchase price for traction substations and the electricity cost for returning three-phase power to the grid, respectively. The price is based on demand. , These are the operation and maintenance costs of energy storage and photovoltaic systems, respectively. , These represent the power purchased from the three-phase power grid and the power transmitted back to the traction substation at time t, respectively. This represents the maximum required power. , These represent the charging and discharging power of the energy storage system at time t, respectively. Let t be the power output of the photovoltaic power generation system at time t; T is the number of scheduling periods in the day-ahead power flow optimization scheduling phase. This refers to the time scale used in the current optimized scheduling phase.

[0061] Taking into account the rapid fluctuation characteristics of train traction load and the scale of the day-ahead power flow optimization scheduling problem, the time scale of the day-ahead optimization scheduling stage is selected as 1 minute, and the scheduling cycle is T=1440.

[0062] Step 3: Based on the flexible traction substation model and the predicted train load and photovoltaic output data of the left and right power supply arms obtained in Step 1, establish the constraints under normal operating conditions of the flexible traction substation with photovoltaic and energy storage systems connected. These constraints include the active power balance constraints of the primary and secondary sides of the traction transformer, the active power balance constraints of the left and right power supply arms, the active power balance constraints of the DC bus in the middle of the back-to-back converters, the active power constraints of the interaction between the primary side common coupling point (PCC) and the three-phase power grid, the active power constraints of the back-to-back converters, the charging and discharging power constraints of the energy storage system, the state of charge constraints of the energy storage system, and the maximum demand constraints. The specific constraints are as follows:

[0063] The traction transformer must maintain a balance of active power between the primary and secondary sides, subject to the following constraints:

[0064]

[0065] In the formula: , Let α and β be the total active power of the power supply arms at time t, respectively.

[0066] The left and right power supply arms must maintain a balance of active power, subject to the following constraints:

[0067]

[0068] In the formula: , These are the active power of the α and β phase converters in the back-to-back converter at time t, respectively. , These represent the total active power loads of the α and β phase power supply arms of the train at time t.

[0069] The DC bus between back-to-back converters must maintain active power balance, subject to the following constraints:

[0070]

[0071] The interactive power at the primary PCC point of the traction substation is constrained by the capacity of the traction transformer, specifically:

[0072]

[0073] In the formula: The power limit for interaction between the traction substation and the three-phase power grid; Let t be a binary variable representing the direction of power interaction between the traction substation and the three-phase power grid. This indicates that the traction substation draws power from the three-phase power grid. This indicates that the traction substation is sending power back to the three-phase power grid.

[0074] The active power compensation of back-to-back converters must meet the rated capacity constraints of the equipment, specifically:

[0075]

[0076] In the formula: The capacity of the i-phase converter equipment.

[0077] The charging and discharging power of the energy storage system must be within the rated charging and discharging power range, with the specific constraints as follows:

[0078]

[0079] In the formula: Rated charge and discharge power for the energy storage system; Let t be a binary variable representing the charging and discharging state of the energy storage system at time t. This indicates that the energy storage system is in a charging state. This indicates that the energy storage system is in a discharging state.

[0080] To prevent overcharging and over-discharging of the energy storage medium, the stored energy must be limited to a certain range. To ensure periodicity during the day-ahead dispatch phase, the energy level of the energy storage system must remain consistent throughout the entire process. The specific constraints that the energy storage system's state of charge must meet are as follows:

[0081]

[0082]

[0083]

[0084] In the formula: Let t be the energy stored in the energy storage system at time t; , These are the charging and discharging efficiencies of the energy storage system, respectively. The state of charge of the energy storage system at time t; This refers to the rated capacity of the energy storage system. , These are the upper and lower limits of the state of charge of the energy storage system, respectively. , These represent the energy levels of the energy storage system at the beginning and end of the scheduling day, respectively.

[0085] In electrified railways, the maximum average load of the traction substation over a 15-minute period is typically used as the maximum demand for the month. Assuming that train operation at the traction substation has a daily periodicity, the maximum demand power on the dispatching day is used as the basic electricity charge for the month. The maximum demand must meet the following constraints:

[0086]

[0087] Step 4: Based on the objective function of the day-ahead power flow optimization scheduling model obtained in Step 2 and the normal operation constraints of the flexible traction substation obtained in Step 3, establish a mixed-integer linear programming model for the day-ahead power flow optimization scheduling stage. Using a commercial programming solver, such as the optimization tool CPLEX developed by IBM, the solver is called through the Yalmip platform in the Matlab environment to solve the mixed-integer linear programming model. This yields the active power plan at the PCC point, the charging and discharging power plan of the energy storage system, and the state of charge plan of the energy storage medium under the goal of global optimal economic operation of the traction substation during the day. The scheduling plan is then issued in advance to the day-ahead operation stage, thus completing the day-ahead power flow optimization scheduling of the flexible traction substation with photovoltaic and energy storage systems connected.

[0088] Step 5: During the intraday rolling optimization operation phase, construct a model predictive control scheme for the flexible traction substation at time k:

[0089] 5.1 Select the vector consisting of the PCC point power (positive value when the traction substation draws power from the three-phase grid, and negative value when the traction substation feeds back power to the three-phase grid) at time k of the flexible traction substation, the charging and discharging power of the energy storage system (positive value when the power is in the discharging state, and negative value when the power is in the charging state), and the state of charge of the energy storage system. The state vector is selected; the increment of the energy storage system output at time k in the flexible traction substation relative to the planned value before the day is chosen. The control vector is formed by selecting the increments of the ultra-short-term predicted power curves of the total train load and photovoltaic output of the left and right power supply arms at time k of the flexible traction substation relative to the day-ahead predicted curve. The disturbance vector is selected; the vector is formed by the power at point PCC of the flexible traction substation at time k and the state of charge of the energy storage system. As the output vector, the prediction model for the predictive control method of the flexible traction substation at time k is established as follows:

[0090]

[0091] In the formula: For the charging and discharging efficiency of the energy storage system; This is the time scale for the intraday rolling optimization phase.

[0092] 5.2 Obtain the ultra-short-term forecast data of traction load and photovoltaic output from time k+1 to time k+P, i.e., within the forecast time domain. The forecast model in 5.1 is iterated step by step within the ultra-short-term forecast time domain to construct the output vector matrix in the forecast time domain as follows:

[0093]

[0094] In the formula: (p=1, ,P) is the output vector of the prediction model at time k, which is composed of the power of the substation PCC point and the state of charge of the energy storage system at time k+p.

[0095] 5.3 Based on the ultra-short-term forecast data of traction load and photovoltaic output obtained in 5.2, the control time domain of the model predictive control scheme is selected from time k+1 to time k+M of the flexible traction substation. To ensure that the intraday operation plan of the traction substation effectively tracks the day-ahead scheduling plan, and to ensure that the deviation between the power at the PCC point and the day-ahead plan is as small as possible, while avoiding excessive power adjustment to the energy storage system, the optimization objective of the model predictive control method at time k is established as follows:

[0096]

[0097] In the formula: The tracking plan target matrix for controlling the output vector in the time domain at time k is given by the power values ​​of the PCC points in the daily plan at each time point in the time domain. (p=1, M), energy storage state of charge (p=1, Composed of (M); The control matrix at time k is the control matrix within the time domain, which is the increment of energy storage power relative to the day-ahead plan at each time point in the time domain. (p=1, The system consists of M and W, which are weighting coefficient matrices. To ensure that the power at the PCC point effectively tracks the day-ahead scheduling plan, the weight of the power at the PCC point must be much greater than the weight of the state of charge of the energy storage system.

[0098] To reduce the impact of traction load and photovoltaic output fluctuations on the PCC point work plan and maximize the economic operation capability of the flexible traction substation, the time scale for the intraday rolling optimization stage is selected as 5 seconds, and the number of rolling optimizations is K=17280; at the same time, the prediction time domain length in the model predictive control is selected as 15 minutes, the number of time periods is P=180, the control time domain length is 3 minutes, and the number of time periods is M=36.

[0099] 5.4 Based on the flexible traction substation model and the ultra-short-term load and photovoltaic output forecast data obtained in 5.2, constraints are established for the substation under normal operating conditions in the control time domain. These constraints include substation power balance constraints, primary-side point of common coupling (PCC) power constraints, back-to-back converter power constraints, energy storage system charging and discharging power constraints, energy storage system state of charge constraints, and maximum demand constraints. The constraints are as follows:

[0100] The traction transformer must maintain a balance of active power between the primary and secondary sides, subject to the following constraints:

[0101]

[0102] In the formula: , These represent the total active power of the α and β phase power supply arms at time p, respectively.

[0103] The left and right power supply arms must maintain a balance of active power, subject to the following constraints:

[0104]

[0105] In the formula: , These are the active power of the α and β phase converters at time p, respectively; , These represent the total active power loads of the α and β phase power supply arms of the train at time p.

[0106] The DC bus between back-to-back converters must maintain active power balance, subject to the following constraints:

[0107]

[0108] The interactive power at the primary PCC point of the traction substation is constrained by the capacity of the traction transformer, specifically:

[0109]

[0110] In the formula: The power exchanged at PCC point is positive when the traction substation purchases power from the three-phase power grid and negative when it is sent back to the three-phase power grid. This is the power limit for interaction between the traction substation and the three-phase power grid.

[0111] The active power compensation of back-to-back converters must meet the rated capacity constraints of the equipment, specifically:

[0112]

[0113] In the formula: The capacity of the i-phase converter equipment.

[0114] The charging and discharging power of the energy storage system must be within the rated charging and discharging power range, with the specific constraints as follows:

[0115]

[0116] In the formula: The power of the energy storage system is represented by positive values ​​for discharging and negative values ​​for charging. The rated charging and discharging power of the energy storage system.

[0117] To prevent overcharging and over-discharging of energy storage media, the energy stored must be limited to a certain range. Specific constraints are as follows:

[0118]

[0119]

[0120] In the formula: Let p be the energy stored in the energy storage system at time p; For the charging and discharging efficiency of the energy storage system; The state of charge of the energy storage system at time p; This refers to the rated capacity of the energy storage system. , These are the upper and lower limits of the state of charge of the energy storage system, respectively. This is the time scale for the intraday rolling optimization phase.

[0121] 5.5 Based on the objective function of the model predictive control scheme obtained in 5.3 and the normal operation constraints of the flexible traction substation obtained in 5.4, a quadratic programming model for the intraday optimized operation stage at time k is established. The model is solved using a quadratic programming solver, such as the quadratic programming problem solver quadprog function provided by Matlab, to obtain the charging and discharging power correction vector of the energy storage system at times k+1, k+2, ..., k+M. The correction amount of the first time period (i.e., from time k+1 to k+2) in this vector is sent to the real-time operation control system of the energy storage system.

[0122] Step 6: In the real-time optimization operation phase, construct a real-time optimization operation method for the flexible traction substation at time k:

[0123] 6.1 Using the model predictive control scheme established at time k-1 of the flexible traction substation in step 5, the correction amount of the charging and discharging power of the energy storage system from time k to time k+1 is obtained. By superimposing this amount on the day-ahead working plan value of the energy storage system from time k to time k+1, the final working plan of the energy storage system from time k to time k+1 can be obtained.

[0124] 6.2 Using a data acquisition device installed in the flexible traction substation, the voltage of the left and right power supply arms is collected in real time from time k to k+1. and Load current of the left and right power supply arms and and the output voltage and current values ​​of the photovoltaic array and Calculate the active power of the total train load on both the left and right power supply arms. and reactive power and photovoltaic power generation Because three-phase power grids adopt a "positive feedback count" or "no feedback count" policy for the active power fed back to electrified railway traction substations, when the photovoltaic output cannot be absorbed by train loads and energy storage systems, it will be directly fed back to the grid through back-to-back converters. This not only fails to bring economic benefits to the operation of flexible traction substations but also occupies additional capacity of the back-to-back converters, reducing their power quality management capabilities and hindering the grid-friendly connection of flexible traction substations. Therefore, during the real-time optimization operation phase, it is necessary to first determine the actual output value of the photovoltaic power generation system. Specific methods include... Figure 3 As shown, based on real-time information collection, it is determined whether there is a phenomenon of PCC point power feeding back to the three-phase power grid at the current moment. If the determination result is that there is no feeding back phenomenon, the photovoltaic grid-connected converter operates in maximum power point tracking mode, controlling the photovoltaic array to output the maximum power that can be generated under the current environmental conditions; otherwise, it adopts power limiting output mode, controlling the photovoltaic array to output the required real-time output power. The real-time output power of the photovoltaic is as follows:

[0125]

[0126] In the formula: The final working plan value for the energy storage system obtained in section 6.1.

[0127] 6.3 Based on the load of the left and right power supply arms obtained in real time and calculated in 6.2, as well as the photovoltaic output power obtained after control by the photovoltaic grid-connected converter, in order to achieve friendly interaction between the traction substation and the three-phase power grid and reduce the negative sequence current on the primary side of the substation, the following objective function for real-time optimization compensation of negative sequence current is established:

[0128]

[0129] In the formula: k T For the traction transformer turns ratio; , These represent the total active power of the α and β phase power supply arms at time n, respectively. , These represent the total reactive power of the α and β phase power supply arms at time n, respectively.

[0130] To meet the real-time control requirements, the time scale for the real-time optimization phase is set to 1 second, and the number of real-time optimization periods N = 86400.

[0131] 6.4 Based on the flexible traction substation model and the real-time load data and photovoltaic control output data obtained in 6.2, constraints are established for the back-to-back converter under normal operating conditions, including power balance constraints on the left and right power supply arms, DC bus power balance constraints, and back-to-back converter capacity constraints. The constraints are as follows:

[0132] The left and right power supply arms must meet the requirements of active power balance and reactive power balance, with the following constraints:

[0133]

[0134] In the formula: , These represent the total active load and total reactive load of the i-phase power supply arm train at time n, respectively. , These represent the active power compensation and reactive power compensation of phase i in the back-to-back converter at time n.

[0135] The DC bus between back-to-back converters must maintain active power balance, subject to the following constraints:

[0136]

[0137] In a back-to-back converter, the compensation power of the left and right converters must meet the equipment capacity constraints, specifically:

[0138]

[0139] 6.5 Based on the objective function for real-time compensation of negative sequence current obtained in 6.3 and the normal operation constraints of the back-to-back converter obtained in 6.4, a nonlinear programming model for the intraday optimized operation phase at time n is established. Using a nonlinear optimization algorithm, such as a sequential quadratic programming algorithm, the model is solved quickly to obtain the real-time active power compensation command of the back-to-back converter at time n. and Real-time reactive power compensation command and The instruction is then sent to the real-time operation control system of the back-to-back converter.

[0140] Step 7: At time k+1, obtain the ultra-short-term forecast data of traction load and photovoltaic output of the traction substation from time k+2 to time k+P+1, establish a model predictive control scheme for the flexible traction substation at time k+1, repeat step 5, establish a quadratic programming model for the intraday optimized operation phase from time k+2 to time k+M+1, solve the model to obtain the charging and discharging power correction vector of the energy storage system at time k+2, k+3, ..., k+M+1, and send the correction amount of the first time period (i.e., from time k+2 to time k+3) in the vector to the real-time operation control system of the energy storage system, and continuously optimize forward in this way;

[0141] Step 8: The correction amount of the energy storage system from time k+1 to time k+2 obtained from the model predictive control scheme of the flexible traction substation at time k is superimposed on the day-ahead working plan value of the energy storage system from time k+1 to time k+2 to obtain the final working plan of the energy storage system from time k+1 to time k+2; repeat step 6 to solve the real-time optimized operation plan from time k+1 to time k+2, that is, to obtain the real-time output power command of photovoltaic, the real-time active power and reactive power compensation command of back-to-back converter from time k+1 to time k+2, and send them to the operation control system of each converter.

[0142] Therefore, by continuously repeating steps 5, 6, 7, and 8 as the daily operating time k increases, the energy management of the flexible traction substation connected to the photovoltaic and energy storage systems can be completed throughout the day.

[0143] Example 2:

[0144] The flexible traction substation structure with photovoltaic and energy storage systems integrated in this invention is as follows: Figure 1 As shown in Table 1, the relevant parameters of some devices in the system are as follows:

[0145] Table 1 System Equipment Parameter Settings

[0146]

[0147] Using a traditional traction power supply system as the reference group and a flexible traction substation with photovoltaic and energy storage systems connected to a single time-scale energy management scheme employing only day-ahead power flow optimization scheduling as the control group, load forecast data for traction substations obtained using a recurrent neural network prediction model were used as day-ahead forecast data, while intraday ultra-short-term forecast data were obtained by superimposing the day-ahead forecast data with the prediction error of a normal distribution. In the reference group of the traditional traction power supply system, the three-phase power grid supplies power to the train load during real-time operation. In the control group employing only day-ahead power flow optimization scheduling, the energy storage system operates according to the day-ahead scheduling plan during real-time operation, while the photovoltaic power generation system outputs power at the real-time maximum power point. The back-to-back converters use the same method as the negative-sequence current optimization compensation step in this invention for negative-sequence compensation, and the fluctuations in train load are smoothed in real-time by the three-phase power grid.

[0148] The comprehensive operating cost of the traction substation obtained from simulation calculations for each scheme is shown in Table 2:

[0149] Table 2 Results of Economic Optimization

[0150]

[0151] As shown in Table 2, flexible traction substations with integrated photovoltaic and energy storage systems can effectively utilize photovoltaic power generation resources and regenerative braking energy, thereby reducing the operating costs of traction substations. Furthermore, when using only the single-time-scale energy management scheme of day-ahead power flow optimization scheduling, the electricity cost of traction substations decreased by 51.2%, the demand cost decreased by 64.5%, and the total operating cost decreased by 48.9%. However, under the multi-time-scale energy management scheme of the present invention, the electricity cost of traction substations decreased by 57.4%, the demand cost decreased by 60.2%, and the total operating cost decreased by 51.2%, demonstrating a further improvement in the economic operation capability of traction substations.

[0152] To evaluate the effectiveness of the multi-timescale energy management scheme in this invention, an evaluation index, mean relative deviation, is defined. This index is used to assess the tracking effect of the PCC point work plan curve after intraday rolling optimization on the day-ahead scheduling plan. A smaller value indicates a better tracking effect on the day-ahead plan; conversely, a larger value indicates a worse tracking effect on the day-ahead plan. The expression for mean relative deviation is as follows:

[0153]

[0154] In the formula: The day-ahead power flow optimization scheduling plan value for the PCC point of the traction substation at time n during the real-time operation phase; This represents the actual value of the PCC point of the traction substation at time n during the real-time operation phase.

[0155] The simulation results for each scheme yielded relevant technical indicators for the traction substation, as shown in Table 3.

[0156] Table 3 Optimization Results of Technical Indicators

[0157]

[0158] As shown in Table 3, when only the day-ahead power flow optimization scheduling scheme is used as the single time scale for energy management, the error between the traction load operation data and photovoltaic output data and the day-ahead forecast values ​​is only smoothed by the three-phase power grid during real-time operation. This results in large power fluctuations at the substation's PCC point and a significant deviation from the day-ahead power flow optimization scheduling. Figure 4As shown. After the multi-timescale energy management scheme of this invention is used to correct the substation's daily operation plan, the adverse effects caused by the prediction errors of traction load and photovoltaic output are effectively reduced. The ability of the traction substation's PCC point to track the day-ahead scheduling plan during daily operation is significantly improved, and the power fluctuation at the substation's PCC point is also lower. Figure 5 As shown.

[0159] Regarding negative sequence current compensation, when using only the day-ahead power flow optimization scheduling as a single time-scale energy management scheme, thanks to the negative sequence compensation capability of the back-to-back converters, the maximum negative sequence current is reduced to 41.43A, and the 95% probability value of the negative sequence current is also reduced to 0.08A compared to the traditional traction power supply system. However, since the real-time photovoltaic output does not adopt a flexible output scheme, the negative sequence current compensation capability of the back-to-back converters is not fully utilized. Therefore, in the real-time operation phase of this invention, a flexible photovoltaic output scheme is considered, further reducing the maximum negative sequence current on the primary side of the traction substation to 30.82A, and the 95% probability value of the negative sequence current is further reduced to 0.06A, resulting in better negative sequence current compensation in the traction substation.

[0160] This invention constructs a DC bus by leading out the intermediate DC link of the back-to-back converter in a flexible traction substation, enabling the integration of photovoltaic power generation and energy storage systems. It employs a multi-timescale energy management scheme, utilizing day-ahead economic optimization scheduling, intraday error rolling correction, and real-time compensation for negative-sequence current, to effectively utilize photovoltaic power generation resources along the railway line and train regenerative braking energy. This reduces the overall operating cost of the traction substation and improves its economic operation capability. Simultaneously, it considers the uncertainties in train traction load and photovoltaic output forecasting, mitigating the adverse effects of day-ahead forecast data errors causing a decrease in the traction substation's energy tracking of the day-ahead plan. Furthermore, it considers optimized compensation for negative-sequence current during real-time operation, facilitating friendly interaction between the traction substation and the three-phase power grid. By improving the economic operation capability of the traction substation and reducing primary-side negative-sequence current through a multi-timescale energy management scheme, it provides an operational plan for the integration of photovoltaic power generation and energy storage systems in electrified railways and lays the foundation for future engineering applications.

[0161] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

[0162] Many other changes and modifications can be made without departing from the concept and scope of this invention. It should be understood that this invention is not limited to the specific embodiments, and the scope of this invention is defined by the appended claims.

Claims

1. An energy management method for flexible traction substations with photovoltaic and energy storage systems integrated, characterized in that, The method divides the system energy management strategy into three time scales, including: S1: In the day-ahead optimization scheduling phase, based on the short-term forecast data of train traction load and photovoltaic output, the objective function is to minimize the comprehensive operating cost of the traction substation, and the constraint condition is to ensure the normal operation of the traction substation. A mixed integer linear programming model for day-ahead power flow optimization scheduling is established. S2: In the intraday rolling optimization stage, the objective function is to minimize the deviation of the power supply plan of the traction substation, and the constraint conditions are constructed by controlling the normal operation of the substation in the control time domain to build a model predictive control scheme for the flexible traction substation. S3: In the real-time optimization operation phase, consider the photovoltaic flexible output scheme and establish a nonlinear programming model with the lowest negative sequence current on the primary side of the traction substation. Solve the real-time power compensation command of the back-to-back converter and construct a real-time optimization operation method for the flexible traction substation. During the real-time optimization operation phase, a real-time optimization operation method for the flexible traction substation at time k is constructed, specifically including: The correction amount of the energy storage system from k to k+1 time obtained by the model predictive control method at k-1 time of flexible traction substation is superimposed with the day-ahead work plan value of the energy storage system from k to k+1 time to obtain the final work plan of the energy storage system from k to k+1 time. Real-time acquisition of the voltage of the left and right power supply arms during time k to k+1. and Load current of the left and right power supply arms and and the output voltage and current values ​​of the photovoltaic array and Calculate the active power of the total train load on both the left and right power supply arms. and reactive power and photovoltaic power generation It then determines whether there is any power feedback from the PCC point to the three-phase grid at the current moment. If there is no feedback, the photovoltaic grid-connected converter operates in maximum power point tracking mode; otherwise, it adopts power-limiting output mode. The real-time output power of the photovoltaic is as follows: ， In the formula: This represents the final planned operating value for the energy storage system. Based on the real-time acquired loads of the left and right power supply arms and the photovoltaic output power obtained after control by the photovoltaic grid-connected converter, the objective function for real-time optimization compensation of negative sequence current is established as follows: ， In the formula: k T For the traction transformer turns ratio; , These represent the total active power of the α and β phase power supply arms at time n, respectively. , These represent the total reactive power of the α and β phase power supply arms at time n, respectively. Based on the flexible traction substation model and the obtained real-time load data and photovoltaic control output data, constraints are established for the back-to-back converter under normal operating conditions, including: power balance constraints of the left and right power supply arms, power balance constraints of the DC bus, and capacity constraints of the back-to-back converter. Based on the objective function for real-time optimization compensation of negative sequence current and the normal operation constraints of the back-to-back converter obtained above, a nonlinear programming model for the intraday optimization operation phase at time n is established. This model is then solved quickly to obtain the real-time active power compensation command of the back-to-back converter at time n. and Real-time reactive power compensation command and The instruction is then sent to the real-time operation control system of the back-to-back converter.

2. The energy management method for flexible traction substations with photovoltaic and energy storage systems connected according to claim 1, characterized in that, Prior to step S1, the method further includes: dividing the energy management method into three time scales: day-ahead, intraday, and real-time.

3. The energy management method for flexible traction substations with photovoltaic and energy storage systems connected according to claim 1, characterized in that, Using the traction load and photovoltaic forecasting method, short-term forecast data of train load and photovoltaic output on the left and right sides of the traction substation during the dispatching day are obtained. The objective function of the mixed-integer linear programming model established by the day-ahead power flow optimization scheduling is as follows: ， In the formula: To schedule the total daily operating cost; , These are the electricity purchase price for traction substations and the electricity cost for returning three-phase power to the grid, respectively. The price is based on demand. , These are the operation and maintenance costs of energy storage and photovoltaic systems, respectively. , These represent the power purchased from the three-phase power grid and the power transmitted back to the traction substation at time t, respectively. This represents the maximum required power. , These represent the charging and discharging power of the energy storage system at time t, respectively. Let t be the power output of the photovoltaic power generation system at time t; T is the number of scheduling periods in the day-ahead power flow optimization scheduling phase. This refers to the time scale used in the current optimized scheduling phase.

4. The energy management method for flexible traction substations with photovoltaic and energy storage systems connected according to claim 1, characterized in that, Based on the flexible traction substation model and short-term forecast data of train load and photovoltaic power output, constraints are established for the day-ahead power flow optimization scheduling stage, including: substation power balance constraints, primary side common coupling point (PCC) power constraints, back-to-back converter capacity constraints, energy storage system charging and discharging power constraints, energy storage medium state of charge constraints, and maximum demand power constraints.

5. The energy management method for flexible traction substations with photovoltaic and energy storage systems connected according to claim 1, characterized in that, After establishing a mixed-integer linear programming model for the day-ahead power flow optimization scheduling phase, the model is solved to obtain the substation PCC point power plan, the energy storage system charging and discharging power plan, and the energy storage medium state of charge plan. This completes the day-ahead power flow optimization scheduling of flexible traction substations with photovoltaic and energy storage system access, and the above scheduling plans are issued in advance to the intraday rolling optimization operation phase.

6. The energy management method for flexible traction substations with photovoltaic and energy storage systems connected according to claim 1, characterized in that, A model predictive control method for time k of the flexible traction substation is established during the intraday rolling optimization operation phase, specifically including: A vector is formed by the PCC point power, energy storage system charging and discharging power, and state of charge at time k in the flexible traction substation. The state vector is selected; the increment of the energy storage system output at time k in the flexible traction substation relative to the planned value before the day is chosen. The control vector is formed by selecting the increments of the ultra-short-term predicted power curves of the total train load and photovoltaic output of the left and right power supply arms at time k of the flexible traction substation relative to the day-ahead predicted curve. The disturbance vector is selected; the vector is formed by the power at point PCC of the flexible traction substation at time k and the state of charge of the energy storage system. As the output vector, a prediction model for the predictive control method of the flexible traction substation at time k is established: ， In the formula: For the charging and discharging efficiency of the energy storage system; This is the time scale for the intraday rolling optimization phase.

7. The energy management method for a flexible traction substation with photovoltaic and energy storage systems connected according to claim 6, characterized in that, At time k, the ultra-short-term forecast data of traction load and photovoltaic output from time k+1 to time k+P are obtained, i.e., within the prediction time domain. The prediction model described above iterates step by step within the ultra-short-term prediction time domain to construct the output vector matrix in the prediction time domain as follows: ， In the formula: (p=1, P) is the output vector of the prediction model at time k, which is composed of the power of the substation PCC point and the state of charge of the energy storage system at time k+p. Based on the acquired ultra-short-term forecast data of traction load and photovoltaic output, the control time domain is selected from time k+1 to time k+M of the flexible traction substation, where the control time domain must be less than or equal to the prediction time domain, i.e., M≤P. The optimization objective of the model predictive control method at time k is as follows: ， In the formula: The tracking plan target matrix for controlling the output vector in the time domain at time k is given by the power values ​​of the PCC points in the daily plan at each time point in the time domain. (p=1, M), energy storage state of charge (p=1, Composed of (M); The control matrix at time k is the control matrix within the time domain, which is the increment of energy storage power relative to the day-ahead plan at each time point in the time domain. (p=1, The matrix consists of M and W, which are weight coefficient matrices.

8. The energy management method for a flexible traction substation with photovoltaic and energy storage systems connected according to claim 7, characterized in that, Based on the flexible traction substation model and the ultra-short-term forecast data of load and photovoltaic output obtained above, constraints are established for the flexible traction substation under normal operating conditions in the control time domain, including: substation power balance constraints, primary side common coupling point power constraints, back-to-back converter capacity constraints, energy storage system charging and discharging power constraints, energy storage medium state of charge constraints, and maximum demand power constraints. Based on the above model prediction control method, the objective function and the normal operation constraints of the flexible traction substation in the control time domain are obtained. A quadratic programming model for the intraday optimized operation stage at time k is established. The quadratic programming model is solved to obtain the charging and discharging power correction vector of the energy storage system at times k+1, k+2, ..., k+M. The correction amount in the first time period of the vector, i.e., from time k+1 to k+2, is sent to the real-time operation control system of the energy storage system.

9. The energy management method for flexible traction substations with photovoltaic and energy storage systems connected according to claim 1, characterized in that, The method further includes: In the ultra-short-term forecast data of traction load and photovoltaic output from k+2 to k+P+1 time, a model predictive control scheme for flexible traction substation at k+1 time is established. The model predictive control method for flexible traction substation at k time is repeatedly established in the intraday rolling optimization operation stage. A quadratic programming model for the intraday optimization operation stage from k+2 to k+M+1 time is established. The model is solved to obtain the charging and discharging power correction vector of energy storage system at k+2, k+3, ..., k+M+1 time. The correction amount in the first time period of the vector, i.e., from k+2 to k+3 time, is sent to the real-time operation control system of energy storage system. This process is continuously rolled forward for optimization. The correction amount of the energy storage system from time k+1 to time k+2 is obtained from the model predictive control scheme of the flexible traction substation at time k, and superimposed with the day-ahead working plan value of the energy storage system from time k+1 to time k+2 to obtain the final working plan of the energy storage system from time k+1 to time k+2. The real-time optimization operation method of the flexible traction substation at time k+1 is constructed by repeating the real-time optimization operation stage, and the real-time optimization operation plan from time k+1 to time k+2 is solved to obtain the real-time output power command of photovoltaic, the real-time active power and reactive power compensation command of back-to-back converters from time k+1 to time k+2, and then sent to the operation control system of each converter. Therefore, as k increases, the process is repeated to complete the energy management method for flexible traction substations that integrate photovoltaic and energy storage systems.

Images