Dynamic energy prediction and distribution method and system for railway power supply safety

Abstract

The invention discloses an energy dynamic prediction distribution method and system for railway power supply safety. Monitoring regenerative braking energy, a traction load state, a residual electric quantity state of an energy storage device and a power grid electricity price state in the railway traction power supply system in real time; constructing a cost function taking power supply safety as a core; establishing a dynamic coupling prediction data model including a traction load, a residual electric quantity state of an energy storage device and a power grid electricity price, and designing a dynamic prediction algorithm considering an energy distribution priority; based on a dynamic prediction algorithm, a distribution coefficient among the bidirectional energy fusion device, the energy storage device and the grid-connected energy feedback device is corrected in a rolling manner; based on the distribution coefficient, an energy distribution path is selected, and dynamic prediction distribution of regenerative braking energy in the railway traction power supply system is completed; according to the method and system provided by the invention, regenerative braking energy can be optimally utilized, and the energy utilization efficiency of the power supply system is remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the field of traction power supply, and particularly relates to an energy dynamic prediction and distribution method and system for railway power supply safety. Background Art

[0002] With the rapid development of China's high-speed railway network and the continuous growth of the operating mileage of electrified railways, the railway traction power supply system is facing increasingly severe energy management challenges. Especially under the conditions of frequent train starts, stops and braking, the system needs to efficiently handle the dynamic distribution problem of regenerative braking energy. Traditional energy management methods mainly adopt fixed threshold control strategies, which are difficult to adapt to complex and changeable operating conditions and are prone to safety hazards such as voltage fluctuations and equipment overload.

[0003] The current energy management of railway traction power supply systems mainly has the following technical bottlenecks: Firstly, the existing systems lack multi-dimensional comprehensive consideration of power supply safety, especially the lack of a safety risk assessment mechanism, making it difficult to achieve a balance among key safety indicators such as voltage stability, current limitation and fault recovery; Secondly, traditional modeling methods are difficult to accurately describe the dynamic characteristics of multi-state coupling of traction loads, energy storage states and grid electricity prices, and it is also difficult to achieve multi-step ahead prediction; More importantly, the existing energy distribution strategies lack predictive active control capabilities, unable to formulate predictive optimal coefficient strategies, and difficult to provide forward-looking control instructions for various energy conversion devices, restricting the system response speed and energy utilization efficiency.

[0004] In terms of safety control, existing technologies often adopt static protection thresholds and cannot adapt to the real-time changes in the system operating state. Especially when dealing with sudden load fluctuations or grid faults, traditional methods are difficult to adjust the energy distribution strategy in a timely manner, which may cause misoperation or refusal to operate of protection devices, seriously affecting power supply reliability. In addition, with the increase in the proportion of new energy grid connection and the deepening of power market reform, the power supply system also needs to take into account the economic operation requirements, which puts higher requirements on the energy dynamic distribution algorithm.

[0005] In view of the above problems, there is an urgent need to develop an energy dynamic prediction and distribution method for railway power supply safety. By establishing a safety cost function for multi-objective optimization, constructing a multi-state dynamic prediction model considering traction load, energy storage state and grid electricity price, and designing a dynamic prediction algorithm with energy distribution priorities, the optimal distribution of regenerative braking energy can be achieved to ensure the safe, stable and economic operation of the power supply system. Summary of the Invention

[0006] In order to solve the problems existing in the background technology, the purpose of the present invention is to provide a method for dynamic prediction and allocation of energy for railway power supply safety. The device for dynamic prediction and allocation of energy includes a two-way energy integration device, an energy storage device, and a grid-connected energy feeding device. The energy storage device uses the acquired energy to power an auxiliary monitoring system, an overhead line intelligent monitoring system, and a station energy management and control system. The method includes the following steps: Step (1): continuously collect and monitor the regenerative braking energy, traction load status, remaining power state SOC of the energy storage device, and power price status of the power grid in the railway traction power supply system; Step (2): constructing a cost function with power supply safety as the core as a constraint condition for determining the energy distribution coefficient between the bidirectional energy integration device, the energy storage device and the grid-connected energy feeding device, wherein the cost function includes factors such as voltage stability, current limitation and fault recovery capability; Step (3): Establish a dynamic coupling prediction data model based on multiple state variables, wherein the state variables include the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state, and the dynamic coupling prediction data model includes a time-varying gradient matrix; based on the dynamic coupling prediction data model, determine a dynamic prediction algorithm that takes into account the energy allocation priority to obtain the energy allocation coefficient between the bidirectional energy financing device, the energy storage device, and the grid-connected energy feeding device: ; in, represents the expansion vector of the forward prediction vector of the energy allocation coefficient, Indicates the energy distribution coefficient of the bidirectional energy transfer device, energy storage device, and grid-connected energy feeding device. The sampling time The step-forward prediction vector, ; is the prediction gain matrix; represents the expansion vector of the forward prediction error difference vector; is the priority gain coefficient; , Represents the identity matrix with 3 rows and 3 columns; is the priority function vector; extract In , , , , They are The allocation coefficients of the bidirectional energy integration device, energy storage device, and grid-connected energy feeding device at the sampling time; Based on the dynamic coupling prediction data model, design an estimation algorithm for the time-varying gradient matrix and the prediction gain matrix; Step (4): Based on the calculation result of the dynamic prediction algorithm, rollingly correct the energy distribution coefficient between the bidirectional energy integration device, the energy storage device and the grid-connected energy feeding device, and select the energy distribution path based on the distribution coefficient; Repeat the above steps (1) to (4) to realize the dynamic prediction and distribution of regenerative braking energy in the railway traction power supply system.

[0007] Furthermore, the dynamic coupling prediction data model based on multi-state variables established in step (3) is: ; in, represents the sampling time; M Represents the prediction step length, which is a positive integer; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. M A vector of step-forward prediction value vectors; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The sampling time m step-forward prediction vector, ; , , They represent the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The sampling time m Step forward prediction value; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. A vector of values ​​at the sampling time, , , They represent the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The value at the sampling time; , represents the identity matrix with 3 rows and 3 columns, is the time-varying gradient matrix, , for The time-varying gradient submatrix at the sampling moment; Represented by vector , , , The vector formed, Indicates the energy distribution coefficient of the bidirectional energy transfer device, energy storage device, and grid-connected energy feeding device. The sampling time step forward prediction increment vector, The energy distribution coefficient is The sampling time Step forward prediction increment value, for ,when When , the energy distribution coefficients of the energy integration device, energy storage device, and grid-connected energy feeding device are The sampling time The step-forward prediction increment value, ; , , Respectively represent the energy distribution coefficients of the bidirectional energy fusion device, energy storage device, and grid-connected energy feeding device. The sampling time Step forward prediction value.

[0008] Furthermore, in step (3), the prediction gain matrix is ​​designed based on the dynamic coupling prediction data model. The estimation algorithm is: ; in, , is a constant, is a one-step forward differentiation, and ; represents one step forward error; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. M The vector of step-forward prediction setpoint vectors, Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The sampling time step-forecast setpoint vector, , , They represent the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The sampling time Step prediction set value; The expansion vector representing the forward prediction error difference vector is composed of the vector , , , , The vector formed, Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The step forward prediction error vector at the sampling moment, denotes the first-order backward difference of the vector ; ; denotes the vector composed of the values of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the sampling moment, , , respectively denote the values of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the sampling moment; denotes the vector composed of the M step forward prediction set value vectors of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state, denotes the step prediction set value vector of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the sampling moment, , , respectively denote the step prediction set values of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the sampling moment.

[0009] Furthermore, in step (3), based on the dynamic coupling prediction data model, the algorithm for designing the estimation of the time-varying gradient matrix includes: When designing the estimation algorithm of the time-varying gradient sub-matrix at the sampling moment, the optimization objective is , where the symbol denotes minimizing the function with respect to the variable taking values, and minimizing the cost function is used as a constraint: ; ; wherein, denotes the first-order backward difference of the vector , denotes the vector composed of the values of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the sampling moment, , , respectively denote the values of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the sampling moment; Indicates the energy distribution coefficient of the bidirectional energy transfer device, energy storage device, and grid-connected energy feeding device. The historical increment vector at the sampling time, The energy distribution coefficient is The historical increment value at the sampling time, for ,when When , the energy distribution coefficients of the energy integration device, energy storage device, and grid-connected energy feeding device are Historical incremental value at the sampling moment; is the step size factor, is the penalty factor, is the two-norm; Design time-varying gradient submatrix , , When the estimation algorithm is: ; in, , Represented by the matrix , , The vector formed, express The time-varying gradient submatrix at the sampling moment, , is a positive integer between 3 and 5; is an intermediate variable, and its calculation formula is: ; , Represented by the matrix , , The vector formed, express The time-varying gradient submatrix at the sampling moment; is a constant.

[0010] Furthermore, the method for selecting the energy distribution path based on the distribution coefficient in step (4) is: like Not 0, path 1 is executed first; if path 1 does not completely consume energy and Not 0, execute path 2; if there is still remaining energy and If it is not 0, execute path 3; Among them, path 1 indicates that the regenerative braking energy is preferentially transferred to the adjacent power supply arm through the bidirectional energy integration device according to the traction load state; path 2 indicates that the remaining regenerative braking energy is allocated to the energy storage device for storage according to the remaining power state SOC adaptive algorithm of the energy storage device; path 3 indicates triggering the grid-connected energy feeding device for reverse feeding.

[0011] Furthermore, in step (2), a cost function with power supply safety as the core is constructed, and the cost function is: ; in, is the cost function coefficient, They respectively represent the voltage stability cost item, the current limiting cost item, and the fault recovery capability cost item.

[0012] Furthermore, the remaining power state SOC adaptive algorithm of the energy storage device in the energy distribution path 2 includes the following steps: When it is detected that the remaining power state SOC of the energy storage device is ≥ the first threshold, reducing the value of the energy storage device allocation coefficient; When the remaining charge state SOC of the energy storage device is less than or equal to a second threshold, energy is prohibited from being released from the energy storage device.

[0013] The energy dynamic prediction and distribution system for railway power supply security includes: Data monitoring module, used to continuously collect and monitor the regenerative braking energy, traction load status, remaining power state SOC of the energy storage device, and power grid electricity price status in the railway traction power supply system; A safety optimization decision module is used to construct a cost function with power supply safety as the core, which is used as a constraint condition for determining the energy distribution coefficient between the bidirectional energy integration device, the energy storage device and the grid-connected energy feeding device, and the cost function includes voltage stability, current limitation, and fault recovery capability factors; A dynamic prediction calculation module is used to establish a dynamic coupling prediction data model based on multiple state variables, wherein the state variables include the traction load state, the remaining power state SOC of the energy storage device, and the power grid price state, and the dynamic coupling prediction data model includes a time-varying gradient matrix; based on the dynamic coupling prediction data model, a dynamic prediction algorithm considering the energy allocation priority is determined to obtain the energy allocation coefficient between the bidirectional energy financing device, the energy storage device, and the grid-connected energy feeding device: ; in, represents the expansion vector of the forward prediction vector of the energy allocation coefficient, Indicates the energy distribution coefficient of the bidirectional energy transfer device, energy storage device, and grid-connected energy feeding device. The sampling time The step-forward prediction vector, ; is the prediction gain matrix; represents the expansion vector of the forward prediction error difference vector; is the priority gain coefficient; , Represents the identity matrix with 3 rows and 3 columns; is the priority function vector; extract In , , , , They are The allocation coefficients of the bidirectional energy integration device, energy storage device, and grid-connected energy feeding device at the sampling time; Based on the dynamic coupling prediction data model, design an estimation algorithm for the time-varying gradient matrix and the prediction gain matrix; A coefficient allocation strategy module, used for rolling correction of the energy allocation coefficient between the bidirectional energy fusion device, the energy storage device and the grid-connected energy feeding device based on the calculation result of the dynamic prediction algorithm; The coefficient allocation execution module is used to select an energy allocation path based on the allocation coefficient.

[0014] Furthermore, the present invention adopts the following technical solutions: A non-transitory computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned method for dynamic energy prediction and allocation for railway power supply security.

[0015] Furthermore, the present invention adopts the following technical solutions: An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the program, the method for dynamic energy prediction and allocation for railway power supply safety as described above is implemented.

[0016] The beneficial technical effects of the present invention are: (1) Based on a real-time updated dynamic coupling prediction data model, the present invention designs a dynamic prediction algorithm that takes into account multi-step prediction errors, calculates the energy allocation coefficients of the bidirectional energy integration device, the energy storage device, and the grid-connected energy feeding device, and accurately predicts the energy supply and demand relationship under different working conditions. In this process, by establishing a cost function for railway power supply safety, voltage stability, current protection, and fault recovery capability are incorporated into a unified optimization framework, thereby realizing dynamic prevention and control of power supply system safety risks and greatly improving the reliability and safety of the traction power supply system. (2) By establishing a dynamic adjustment mechanism with energy distribution priorities, the present invention intelligently coordinates various energy conversion methods such as two-way energy flow-through, energy storage, and grid-connected power feeding, enabling the optimal utilization of regenerative braking energy and significantly improving the energy utilization efficiency of the power supply system. (3) The dynamic prediction algorithm considering energy distribution priorities proposed by the present invention does not rely on an accurate mathematical model. This method has obvious advantages when dealing with a traction power supply system with strong time-varying, strong non-linear, and strong coupling characteristics, and is easy to be applied in practice. Brief Description of the Drawings

[0017] Figure 1 It is a schematic flow chart of the energy dynamic prediction and distribution method for railway power supply safety provided in Embodiment 1 of the present invention; Figure 2 It is a schematic diagram of the module connection of the energy dynamic prediction and distribution system for railway power supply safety provided in Embodiment 2 of the present invention; Figure 3 It is a power supply system of the energy storage device provided in Embodiment 1 of the present invention; Figure 4 It is a schematic diagram of the multi-purpose power supply method of the energy storage device provided in Embodiment 1 of the present invention. Detailed Embodiment

[0018] The present invention discloses an energy dynamic prediction and distribution method and system for railway power supply safety; it real-time monitors the regenerative braking energy, traction load status, remaining battery capacity status SOC of the energy storage device, and grid electricity price status in the railway traction power supply system; constructs a cost function with power supply safety as the core; establishes a dynamic coupling prediction data model including the traction load, remaining battery capacity status SOC of the energy storage device, and grid electricity price, and designs a dynamic prediction algorithm considering energy distribution priorities; based on the dynamic prediction algorithm, rolls and corrects the distribution coefficients among the two-way energy flow-through device, energy storage device, and grid-connected power feeding device; based on the distribution coefficients, selects the energy distribution path to complete the dynamic prediction and distribution of the regenerative braking energy in the railway traction power supply system; the method and system provided by the present invention can predict the energy supply and demand relationship under different working conditions, enable the optimal utilization of the regenerative braking energy, and significantly improve the energy utilization efficiency of the power supply system.

[0019] The following further clearly and completely describes the energy dynamic prediction and distribution method and system for railway power supply safety provided by the present invention with reference to the drawings: Embodiment 1

[0020] Figure 1A schematic flow chart of a method for dynamic energy prediction and allocation for railway power supply security provided by this embodiment is given. The method for dynamic energy prediction and allocation for railway power supply security provided by this embodiment, wherein the device for dynamic energy prediction and allocation includes a bidirectional energy fusion device, an energy storage device, and a grid-connected energy feeding device, and the method includes the following steps: Step (1): continuously collect and monitor the regenerative braking energy, traction load status, remaining power state SOC of the energy storage device, and power price status of the power grid in the railway traction power supply system; Step (2): constructing a cost function with power supply safety as the core as a constraint condition for determining the energy distribution coefficient between the bidirectional energy integration device, the energy storage device and the grid-connected energy feeding device, wherein the cost function includes factors such as voltage stability, current limitation and fault recovery capability; Step (3): Establish a dynamic coupling prediction data model based on multiple state variables, wherein the state variables include the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state, and the dynamic coupling prediction data model includes a time-varying gradient matrix; based on the dynamic coupling prediction data model, design a dynamic prediction algorithm that considers the energy allocation priority to obtain the energy allocation coefficient between the bidirectional energy financing device, the energy storage device, and the grid-connected energy feeding device: ; in, The expansion vector representing the forward prediction vector of the energy allocation coefficient is composed of the vector , , The vector formed, Indicates the energy distribution coefficient of the bidirectional energy transfer device, energy storage device, and grid-connected energy feeding device. The sampling time The step-forward prediction vector, ; is the prediction gain matrix; The expansion vector representing the forward prediction error difference vector is composed of the vector , , , , The vector formed, Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The sampling time The step-forward prediction error vector, Representation vector The first-order backward difference of , is an adjustable time window parameter; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid M The vector of step-forward prediction setpoint vectors, Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The sampling time step-forecast setpoint vector, , , They represent the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The sampling time Step prediction set value; is the priority gain coefficient; , Represents the identity matrix with 3 rows and 3 columns; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. A vector of values ​​at the sampling time, , , They represent the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The value at the sampling time; is the priority function vector, represents the priority vector, , , Respectively represent the priorities of the bidirectional energy integration device, energy storage device, and grid-connected energy feeding device; extract In , , , , They are The allocation coefficients of the bidirectional energy integration device, energy storage device, and grid-connected energy feeding device at the sampling time; Based on the dynamic coupling prediction data model, design an estimation algorithm for the time-varying gradient matrix and the prediction gain matrix; Step (4): Based on the calculation result of the dynamic prediction algorithm, the energy distribution coefficient between the bidirectional energy fusion device, the energy storage device and the grid-connected energy feeding device is revised in a rolling manner, and the energy distribution path is selected based on the distribution coefficient; if Not 0, path 1 is executed first; if path 1 does not completely consume energy and Not 0, execute path 2; if there is still remaining energy and If it is not 0, execute path 3; Repeat the above steps (1) to (4) to realize the dynamic prediction and distribution of regenerative braking energy in the railway traction power supply system.

[0021] In detail, based on the allocation coefficient, the method for selecting the energy allocation path is as follows: the energy allocation coefficient between the bidirectional energy fusion device, the energy storage device and the grid-connected energy feeding device refers to the vector calculated by the dynamic prediction algorithm ,in , , They are The allocation coefficients of the bidirectional energy integration device, energy storage device, and grid-connected energy feeding device at the sampling time are as follows: Not 0, path 1 is executed first; if path 1 does not completely consume energy and Not 0, execute path 2; if there is still remaining energy and If it is not 0, execute path 3; path 1 indicates that the regenerative braking energy is preferentially transferred to the adjacent power supply arm through the bidirectional energy transfer device according to the traction load state; path 2 indicates that the remaining regenerative braking energy is allocated to the energy storage device for storage according to the remaining power state SOC adaptive algorithm of the energy storage device; path 3 indicates triggering the grid-connected energy feeding device for reverse feeding.

[0022] In step (2), a cost function with power supply safety as the core is constructed, and the cost function is: ; in, is the cost function coefficient, Respectively represent the voltage stability cost item, the current limiting cost item, and the fault recovery capability cost item; The voltage stability cost term is used to penalize the traction network voltage from deviating from the target voltage value: ; in, Indicates the traction network i The actual voltage of the node, Indicates the target voltage value of the traction network. Indicates the number of traction network nodes; The current limit penalty is used to punish the traction network current exceeding the rated current value: ; in, Indicates the traction network i The actual current of each node, Indicates the rated current value; The fault recovery capability cost item is: ; in, Indicates the traction power supply system l The recovery time at each fault, Indicates l The influence coefficient of each fault location, Indicates the number of faults in the traction power supply system.

[0023] In addition, the method of establishing the dynamic coupling prediction data model based on multi-state variables in step (3) is: ; in, represents the sampling time; M Represents the prediction step length, which is a positive integer; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. M A vector of step-forward prediction value vectors; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The sampling time m step-forward prediction vector, ; , , They represent the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The sampling time m Step forward prediction value; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. A vector of values ​​at the sampling time, , , They represent the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The value at the sampling time; , represents the identity matrix with 3 rows and 3 columns, is the time-varying gradient matrix, , for The time-varying gradient submatrix at the sampling moment; Represented by vector , , , The vector formed, Indicates the energy distribution coefficient of the bidirectional energy transfer device, energy storage device, and grid-connected energy feeding device. The sampling time step forward prediction increment vector, The energy distribution coefficient is The sampling time Step forward prediction increment value, for ,when When , the energy distribution coefficients of the energy integration device, energy storage device, and grid-connected energy feeding device are The sampling time The step-forward prediction increment value, ; , , Respectively represent the energy distribution coefficients of the bidirectional energy fusion device, energy storage device, and grid-connected energy feeding device. The sampling time Step forward prediction value; In step (3), based on the dynamic coupling prediction data model, a method for designing a dynamic prediction algorithm considering energy allocation priority is as follows: ; in, The expansion vector representing the forward prediction vector of the energy allocation coefficient is composed of the vector , , The vector formed, Indicates the energy distribution coefficient of the bidirectional energy transfer device, energy storage device, and grid-connected energy feeding device. The sampling time The step-forward prediction vector, ; is the prediction gain matrix; The expansion vector representing the forward prediction error difference vector is composed of the vector , , , , The vector formed, Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The sampling time The step-forward prediction error vector, Representation vector The first-order backward difference of , It is an adjustable time window parameter, which is generally a positive integer within 5; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid MThe vector formed by the step-ahead prediction setpoint vectors represents the traction load status, the state of charge SOC of the energy storage device, and the grid electricity price status at the sampling time for the step-ahead prediction setpoint vector, , , respectively representing the traction load status, the state of charge SOC of the energy storage device, and the grid electricity price status at the sampling time for the step-ahead prediction setpoint; is the priority gain coefficient; , represents the 3x3 identity matrix; represents the vector formed by the values of the traction load status, the state of charge SOC of the energy storage device, and the grid electricity price status at the sampling time, , , respectively representing the values of the traction load status, the state of charge SOC of the energy storage device, and the grid electricity price status at the sampling time; is the priority function vector, represents the priority vector, , , respectively representing the priorities of the bidirectional energy integration device, the energy storage device, and the grid-connected power feeding device; Extract from , where the , , , are respectively the distribution coefficients of the bidirectional energy integration device, the energy storage device, and the grid-connected power feeding device at the sampling time.

[0024] Furthermore, in step (3), the estimation algorithms for designing the time-varying gradient matrix and the prediction gain matrix based on the dynamic coupling prediction data model include: When designing the estimation algorithm for the time-varying gradient sub-matrix at the sampling time, the optimization objective is , where the symbol represents minimizing the function with respect to the variable taking values. In this process, the cost function is minimized as a constraint: ; ; Among them, Representation vector The first-order backward difference of Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. A vector of values ​​at the sampling time, , , They represent the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The value at the sampling time; Indicates the energy distribution coefficient of the bidirectional energy transfer device, energy storage device, and grid-connected energy feeding device. The historical increment vector at the sampling time, The energy distribution coefficient is The historical increment value at the sampling time, for ,when When , the energy distribution coefficients of the energy integration device, energy storage device, and grid-connected energy feeding device are Historical incremental value at the sampling moment; is the step size factor, is the penalty factor, is the two-norm; Design time-varying gradient submatrix , , When the estimation algorithm is used, ; in, , Represented by the matrix , , The vector formed, express The time-varying gradient submatrix at the sampling moment, , is a positive integer between 3 and 5; is an intermediate variable, and its calculation formula is: ; , Represented by the matrix , , The vector formed, express The time-varying gradient submatrix at the sampling moment; is a constant; Based on the dynamic coupling prediction data model, the prediction gain matrix is ​​designed When estimating the algorithm, ; Among them, , is a constant, is the one-step forward differential, and ; represents the one-step forward error; represents the dynamic coupling prediction data model in step (3); represents the M vector composed of the one-step forward prediction set value vectors of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state, represents the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the step prediction set value vector at the sampling moment, , , respectively represent the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the step prediction set value at the sampling moment; represents the M vector composed of the one-step forward prediction value vectors of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state, represents the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the step prediction value vector at the sampling moment, , , respectively represent the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the step prediction value at the sampling moment; represents the value of the prediction gain matrix at the sampling moment; represents the expansion vector of the forward prediction error difference vector, which is composed of the vectors , , , , constitute the vector, represents the one-step forward prediction error vector of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the step at the sampling moment, represents the vector The first-order backward difference of ; , Represents the identity matrix with 3 rows and 3 columns; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. A vector of values ​​at the sampling time, , , They represent the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The value at the sampling time; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid M The vector of step-forward prediction setpoint vectors, Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The sampling time step-forecast setpoint vector, , , They represent the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The sampling time Step prediction set value; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. A vector of values ​​at the sampling time, , , They represent the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The value at the sampling time; , represents the identity matrix with 3 rows and 3 columns, is the time-varying gradient matrix, , for The time-varying gradient submatrix at the sampling moment; Represented by vector , , , The vector formed, Indicates the energy distribution coefficient of the bidirectional energy transfer device, energy storage device, and grid-connected energy feeding device. The sampling time step forward prediction increment vector, The energy distribution coefficient is The sampling time Step forward prediction increment value, for ,when When , the energy distribution coefficients of the energy integration device, energy storage device, and grid-connected energy feeding device are The sampling time The step-forward prediction increment value, ; , , Respectively represent the energy distribution coefficients of the bidirectional energy fusion device, energy storage device, and grid-connected energy feeding device. The sampling time Step forward prediction value.

[0025] The remaining charge state SOC adaptive algorithm of the energy storage device in the energy distribution path 2 includes the following steps: When it is detected that the remaining power state SOC of the energy storage device is greater than or equal to the first threshold, the energy storage device allocation coefficient is reduced. The first threshold is used to prevent energy storage from being overcharged, and its value is based on the safety window parameter calibrated by the energy storage device; When the remaining power state SOC of the energy storage device is less than or equal to a second threshold, energy is prohibited from being released from the energy storage device; the second threshold is used to avoid over-discharge, and its value is based on a safety window parameter calibrated by the energy storage device.

[0026] In detail, the priority function vector The definition includes the following steps: First, assign priorities to bidirectional energy transfer devices, energy storage devices, and grid-connected energy feeding devices. , , 1 is the highest priority, 2 is the medium priority, and 3 is the lowest priority; Secondly, define the priority functions for the bidirectional energy transfer device, energy storage device, and grid-connected energy feeding device respectively. ,in , used to indicate that a higher priority gets a larger priority function value; Finally, the priority function vector is formed ; In step (4), the energy distribution coefficient between the bidirectional energy fusion device, the energy storage device and the grid-connected energy feeding device is revised based on the dynamic prediction algorithm. The rolling revision refers to the real-time update of the dynamic prediction algorithm. The prediction gain matrix in ,vector and the priority function vector .

[0027] It should be noted that after the energy storage device obtains energy, its power supply is used in many aspects. Figure 3 This is the power supply system of the energy storage device provided in this embodiment. Figure 4 This is a schematic diagram of the multi-purpose power supply method of the energy storage device provided in this embodiment. The specific multi-purpose power supply method includes the following steps: (a) Set the power supply start threshold ESS_start and the power supply switching threshold ESS_switch of the energy storage device, where ESS_switch > ESS_start; monitor the state of charge (SOC) of the energy storage device. (b) When the state of charge (SOC) of the energy storage device ≥ ESS_start, start power supply in the following priority order: Highest priority: Auxiliary monitoring system, including environmental monitoring subsystem and security prevention subsystem; Secondary priority: Catenary intelligent monitoring system, including monitoring units for electrical connection clamps, insulators, and positioning devices; Third priority: Station yard energy management and control system, including electronic systems for heating, ventilation, and air conditioning and commercial electronic systems; Among them, for the power supply objects with secondary priority and third priority, dynamic coefficient allocation is implemented according to the power supply distance, and the electrical equipment with a distance from the energy storage device ≤ the set radius is given priority to obtain power supply; If the state of charge (SOC) of the energy storage device is less than ESS_start, return to step (a) and continuously monitor the state of charge (SOC) of the energy storage device; (c) When the state of charge (SOC) of the energy storage device ≥ ESS_switch, automatically switch back to the original power supply mode; if the state of charge (SOC) of the energy storage device is less than ESS_switch, return to step (b) and judge the magnitude relationship between the state of charge (SOC) of the energy storage device and ESS_start. Embodiment 2

[0028] Figure 2 This is a schematic diagram of the module connection of the energy dynamic prediction and allocation system for railway power supply safety provided in this embodiment. The energy dynamic prediction and allocation system for railway power supply safety provided in this embodiment includes: A data monitoring module, which is used to continuously collect and monitor the regenerative braking energy, traction load state, state of charge (SOC) of the energy storage device, and grid electricity price state in the railway traction power supply system; A safety optimization decision module is used to construct a cost function with power supply safety as the core, which is used as a constraint condition for determining the energy distribution coefficient between the bidirectional energy integration device, the energy storage device and the grid-connected energy feeding device, and the cost function includes voltage stability, current limitation, and fault recovery capability factors; A dynamic prediction calculation module is used to establish a dynamic coupling prediction data model based on multiple state variables, wherein the state variables include the traction load state, the remaining power state SOC of the energy storage device, and the power grid price state, and the dynamic coupling prediction data model includes a time-varying gradient matrix; based on the dynamic coupling prediction data model, a dynamic prediction algorithm considering the energy allocation priority is determined to obtain the energy allocation coefficient between the bidirectional energy financing device, the energy storage device, and the grid-connected energy feeding device: ; in, represents the expansion vector of the forward prediction vector of the energy allocation coefficient, Indicates the energy distribution coefficient of the bidirectional energy transfer device, energy storage device, and grid-connected energy feeding device. The sampling time The step-forward prediction vector, ; is the prediction gain matrix; represents the expansion vector of the forward prediction error difference vector; is the priority gain coefficient; , Represents the identity matrix with 3 rows and 3 columns; is the priority function vector; extract In , , , , They are The allocation coefficients of the bidirectional energy integration device, energy storage device, and grid-connected energy feeding device at the sampling time; Based on the dynamic coupling prediction data model, design an estimation algorithm for the time-varying gradient matrix and the prediction gain matrix; A coefficient allocation strategy module, used for rolling correction of the energy allocation coefficient between the bidirectional energy fusion device, the energy storage device and the grid-connected energy feeding device based on the calculation result of the dynamic prediction algorithm; A coefficient allocation execution module, used for selecting an energy allocation path based on the allocation coefficient; Furthermore, the present invention adopts the following technical solutions: A non-transitory computer-readable storage medium stores a computer program thereon. When the computer program is executed by a processor, it implements the energy dynamic prediction and allocation method for railway power supply safety as described above.

[0029] Furthermore, the present invention adopts the following technical solutions: An electronic device includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the energy dynamic prediction and allocation method for railway power supply safety as described above.

[0030] Through the description of the above embodiments, those skilled in the art can clearly understand that the facilities of the present invention can be implemented by means of software plus a necessary general hardware platform. The embodiments of the present invention can be implemented using existing processors, or by a dedicated processor used for this purpose or other purposes in a suitable system, or by a hardwired system. The embodiments of the present invention also include a non-transitory computer-readable storage medium, which includes a machine-readable medium for carrying or having machine-executable instructions or data structures stored thereon; such a machine-readable medium can be any available medium accessible by a general or special computer or other machine with a processor. For example, such a machine-readable medium can include RAM, ROM, EPROM, EEPROM, CD-ROM or other optical disc memories, magnetic disk memories or other magnetic storage devices, or any other medium that can be used to carry or store the required program code in the form of machine-executable instructions or data structures and can be accessed by a general or special computer or other machine with a processor. When information is transmitted or provided to a machine through a network or other communication connection (hardwired, wireless, or a combination of hardwired and wireless), this connection is also regarded as a machine-readable medium.

[0031] So far, the technical solutions of the present invention have been described in combination with the preferred embodiments shown in the drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the protection scope of the present invention.

Claims

1. Energy dynamic prediction and allocation method for railway power supply safety, characterized in that The device for dynamic energy prediction and allocation includes a bidirectional energy integration device, an energy storage device, and a grid-connected energy feeding device. The energy storage device uses the acquired energy to power the auxiliary monitoring system, the contact network intelligent monitoring system, and the station energy management and control system; The method comprises the following steps: Step (1): continuously collect and monitor the regenerative braking energy, traction load status, remaining power state SOC of the energy storage device, and power price status of the power grid in the railway traction power supply system; Step (2): constructing a cost function with power supply safety as the core as a constraint condition for determining the energy distribution coefficient between the bidirectional energy integration device, the energy storage device and the grid-connected energy feeding device, wherein the cost function includes factors such as voltage stability, current limitation and fault recovery capability; Step (3): Establish a dynamic coupling prediction data model based on multiple state variables, wherein the state variables include the traction load state, the remaining power state SOC of the energy storage device, and the power grid price state, and the dynamic coupling prediction data model includes a time-varying gradient matrix; based on the dynamic coupling prediction data model, determine a dynamic prediction algorithm that considers the energy allocation priority; based on the dynamic coupling prediction data model, design an estimation algorithm for the time-varying gradient matrix and the prediction gain matrix; Among them, the dynamic prediction algorithm is: ； Among them, represents the expansion vector of the forward prediction vector of the energy distribution coefficient, represents the energy distribution coefficients of the bi-directional energy integration device, the energy storage device, and the grid-connected power feeding device at the step forward prediction vector at the sampling moment, ; is the prediction gain matrix; represents the expansion vector of the forward prediction error difference vector; is the priority gain coefficient; , represents the 3×3 identity matrix; is the priority function vector; extract In , , , , They are The allocation coefficients of the bidirectional energy integration device, energy storage device, and grid-connected energy feeding device at the sampling time; Step (4): Based on the calculation result of the dynamic prediction algorithm, the energy distribution coefficient between the bidirectional energy integration device, the energy storage device and the grid-connected energy feeding device is revised in a rolling manner, and the energy distribution path is selected based on the distribution coefficient.

2. The energy dynamic prediction and allocation method for railway power supply safety according to claim 1, wherein The dynamic coupling prediction data model based on multi-state variables established in step (3) is: ； in, represents the sampling time; M Represents the prediction step length, which is a positive integer; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. M A vector of step-forward prediction value vectors; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The sampling time m step-forward prediction vector, ; , , They represent the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The sampling time m Step forward prediction value; Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. A vector of values ​​at the sampling time, , , They represent the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The value at the sampling time; , represents the identity matrix with 3 rows and 3 columns, is the time-varying gradient matrix, , for The time-varying gradient submatrix at the sampling moment; Represented by vector , , , The vector formed, Indicates the energy distribution coefficient of the bidirectional energy transfer device, energy storage device, and grid-connected energy feeding device. The sampling time step-forward prediction increment vector, The energy distribution coefficient is The sampling time Step forward prediction increment value, for ,when When , the energy distribution coefficients of the energy integration device, energy storage device, and grid-connected energy feeding device are The sampling time The step-forward prediction increment value, ; , , Respectively represent the energy distribution coefficients of the bidirectional energy fusion device, energy storage device, and grid-connected energy feeding device. The sampling time Step forward prediction value.

3. The energy dynamic prediction and allocation method for railway power supply safety according to claim 2, characterized in that In step (3), based on the dynamic coupling prediction data model, design the prediction gain matrix The estimation algorithm of is as follows: ； Among them, , is a constant, is a one-step forward differential, and ; representing a one-step forward error; A vector composed of the traction load state, the state of charge (SOC) of the energy storage device, and the grid electricity price state M of the step-ahead prediction setpoint vector, indicating the traction load state, the state of charge (SOC) of the energy storage device, and the grid electricity price state at the sampling moment of the step prediction setpoint vector, , , respectively indicating the traction load state, the state of charge (SOC) of the energy storage device, and the grid electricity price state at the sampling moment of the step prediction setpoint; denotes the expansion vector of the forward prediction error difference vector, which is composed of vectors , , , , ; denotes the forward prediction error vector of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the th step at the sampling moment, ; denotes the first-order backward difference of the vector ; ; denotes the vector composed of the values of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the sampling moment, , , respectively denote the values of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the sampling moment; denotes the vector composed of the M th step forward prediction set value vectors of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state, denotes the th step prediction set value vector of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the sampling moment, , , respectively denote the th step prediction set values of the traction load state, the remaining power state SOC of the energy storage device, and the grid electricity price state at the sampling moment.

4. The energy dynamic prediction and allocation method for railway power supply safety according to claim 2, characterized in that In step (3), based on the dynamic coupling prediction data model, designing an estimation algorithm for the time-varying gradient matrix includes: Design Time-varying gradient sub-matrix at the sampling moment When estimating the algorithm of , where the symbol represents minimizing the function with respect to the variable taking values, and minimizing the cost function is used as a constraint: ； in, Representation vector The first-order backward difference of Indicates the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. A vector of values ​​at the sampling time, , , They represent the traction load status, the remaining power state SOC of the energy storage device, and the power price status of the grid. The value at the sampling time; Indicates the energy distribution coefficient of the bidirectional energy transfer device, energy storage device, and grid-connected energy feeding device. The historical increment vector at the sampling time, The energy distribution coefficient is The historical increment value at the sampling time, for ,when When , the energy distribution coefficients of the energy integration device, energy storage device, and grid-connected energy feeding device are Historical incremental value at the sampling moment; is the step size factor, is the penalty factor, is the two-norm; Designing a time-varying gradient submatrix , , when estimating the algorithm of ； Among them, , denotes the vector formed by the matrix , , ; denotes the time-varying gradient submatrix at the sampling moment, , is a positive integer between 3 and 5; is an intermediate variable, and its calculation formula is: ； , denotes the vector formed by the matrix , , ; denotes the time-varying gradient sub-matrix at the sampling moment; is a constant.

5. The energy dynamic prediction and distribution method for railway power supply safety according to claim 1, wherein The method for selecting the energy distribution path based on the distribution coefficient in step (4) is: If is not 0, path 1 is preferentially executed; if path 1 does not fully absorb the energy and is not 0, path 2 is executed; If there is still remaining energy and is not zero, execute Path 3; Among them, path 1 indicates that the regenerative braking energy is preferentially transferred to the adjacent power supply arm through the bidirectional energy integration device according to the traction load state; path 2 indicates that the remaining regenerative braking energy is allocated to the energy storage device for storage according to the remaining power state SOC adaptive algorithm of the energy storage device; path 3 indicates triggering the grid-connected energy feeding device for reverse feeding.

6. The energy dynamic prediction and allocation method for railway power supply safety according to claim 1, wherein In step (2), a cost function with power supply safety as the core is constructed, and the cost function is: ； Among them, is the cost function coefficient, respectively representing the voltage stability cost item, the current limit cost item, and the fault recovery ability cost item.

7. The energy dynamic prediction and allocation method for railway power supply safety according to claim 5, characterized in that, The remaining charge state SOC adaptive algorithm of the energy storage device in the energy distribution path 2 includes the following steps: When it is detected that the remaining power state SOC of the energy storage device is ≥ the first threshold, reducing the value of the energy storage device allocation coefficient; When the remaining charge state SOC of the energy storage device is less than or equal to a second threshold, energy is prohibited from being released from the energy storage device.

8. Energy dynamic prediction and distribution system for railway power supply safety, characterized in that, include: Data monitoring module, used to continuously collect and monitor the regenerative braking energy, traction load status, remaining power state SOC of the energy storage device, and power grid electricity price status in the railway traction power supply system; A safety optimization decision module is used to construct a cost function with power supply safety as the core, which is used as a constraint condition for determining the energy distribution coefficient between the bidirectional energy integration device, the energy storage device and the grid-connected energy feeding device, and the cost function includes voltage stability, current limitation, and fault recovery capability factors; A dynamic prediction calculation module is used to establish a dynamic coupling prediction data model based on multiple state variables, wherein the state variables include the traction load state, the remaining power state SOC of the energy storage device, and the power grid price state, and the dynamic coupling prediction data model includes a time-varying gradient matrix; based on the dynamic coupling prediction data model, a dynamic prediction algorithm considering the energy allocation priority is determined to obtain the energy allocation coefficient between the bidirectional energy financing device, the energy storage device, and the grid-connected energy feeding device: ； Among them, represents the expansion vector of the forward prediction vector of the energy distribution coefficient, represents the energy distribution coefficients of the bi-directional energy integration device, the energy storage device, and the grid-connected energy feeding device at the sampling time at the step forward prediction vector, ; is the prediction gain matrix; represents the expansion vector of the forward prediction error difference vector; is the priority gain coefficient; , represents the 3×3 identity matrix; is the priority function vector; extract In , , , , They are The allocation coefficients of the bidirectional energy integration device, energy storage device, and grid-connected energy feeding device at the sampling time; Based on the dynamic coupling prediction data model, design an estimation algorithm for the time-varying gradient matrix and the prediction gain matrix; A coefficient allocation strategy module, used for rolling correction of the energy allocation coefficient between the bidirectional energy fusion device, the energy storage device and the grid-connected energy feeding device based on the calculation result of the dynamic prediction algorithm; The coefficient allocation execution module is used to select an energy allocation path based on the allocation coefficient.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, the method for dynamic energy prediction and allocation for railway power supply security as described in any one of claims 1 to 7 is implemented.

10. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, the method for dynamic energy prediction and allocation for railway power supply security as described in any one of claims 1 to 7 is implemented.

Images