Method for evaluating energy saving quantity after micro-grid with light storage is connected to electrified railway and related equipment

Abstract

A method for evaluating energy-saving quantity of a photovoltaic and energy storage micro-grid connected to an electrified railway and related equipment. In the method, an evaluation boundary is determined, and basic parameters are set; multi-source basic data is collected in an evaluation period according to a time step, and actual grid power consumption is calculated; a photovoltaic and energy storage collaborative energy-saving quantity evaluation model is constructed; photovoltaic energy-saving quantity is calculated through a photovoltaic energy-saving quantity calculation sub-model; energy storage peak-valley arbitrage energy-saving quantity and energy storage grid loss optimization energy-saving quantity are calculated through an energy storage energy-saving quantity calculation sub-model; traction transformer loss optimization energy-saving quantity and total energy-saving quantity are calculated through a total energy-saving quantity calculation sub-model; a precision index is calculated, and an evaluation result is obtained when the precision index is less than or equal to a preset precision threshold; and the photovoltaic and energy storage micro-grid is adjusted according to the evaluation result. By implementing the technical solution provided in the application, the energy-saving benefits of each link are quantified relatively accurately, and a basis is provided for railway operators to carry out fine energy-saving benefit accounting.

Description

Technical Field

[0001] This application relates to the field of energy-saving assessment technology for electrified railways, specifically to a method and related equipment for assessing energy savings after a photovoltaic-storage microgrid is connected to an electrified railway. Background Technology

[0002] Electrified railways consume a large amount of electricity during operation. Integrating a photovoltaic (PV) power generation and energy storage microgrid into the traction power supply system has become an application method to reduce electricity costs. To calculate the energy-saving benefits after the PV-storage microgrid is connected, the current practice is to record grid electricity metering data for a period of time before and after the connection, and use the difference in grid electricity consumption before and after the connection as the basis for estimating the energy savings.

[0003] However, the above approach has a prominent problem: the traction load fluctuates greatly, and the difference in grid power consumption before and after connection is mixed with interference factors such as load fluctuation and seasonal differences. It is impossible to separate various energy-saving contributions from the total amount, resulting in low accuracy of energy-saving assessment results, which is difficult to meet the actual needs of railway operators to conduct refined accounting of energy-saving benefits. Summary of the Invention

[0004] To address the aforementioned technical issues, this application provides a method and related equipment for assessing energy savings after a photovoltaic-storage microgrid is connected to an electrified railway.

[0005] The first aspect of this application provides a method for assessing energy savings after a photovoltaic-storage microgrid is connected to an electrified railway, employing the following technical solution: The evaluation boundary is determined, and basic parameters are set based on the evaluation boundary. The basic parameters include time step, evaluation period, traction transformer loss coefficient, energy storage charging and discharging efficiency, and time period division results of railway grid electricity price. According to the time step, multi-source basic data is collected within the evaluation period. The multi-source basic data includes traction load value, photovoltaic output value, energy storage charging power, energy storage discharging power and grid power metering value at each time. The grid power metering values ​​at each time within the evaluation period are accumulated to obtain the actual grid power consumption. Based on the multi-source basic data, the loss coefficient of the traction transformer, the energy storage charging and discharging efficiency, and the time period division results, a photovoltaic-storage synergistic energy-saving assessment model is constructed. The photovoltaic-storage synergistic energy-saving assessment model includes a photovoltaic energy-saving calculation sub-model, an energy storage energy-saving calculation sub-model, and a total energy-saving calculation sub-model. The photovoltaic energy saving calculation sub-model calculates the photovoltaic energy saving based on the photovoltaic output value, traction load value, and energy storage discharge power at each time. The energy-saving calculation sub-model for energy storage calculates the energy-saving energy from peak-valley arbitrage and the energy-saving energy from energy storage network loss optimization based on the multi-source basic data, the energy storage charging and discharging efficiency, and the time period division results. Using the total energy saving calculation sub-model, the energy saving of photovoltaic power, the energy saving of energy storage peak-valley arbitrage, the energy saving of energy storage network loss optimization, the loss coefficient of traction transformer, the traction load value at each time, the photovoltaic output value at each time, the energy storage discharge power at each time, and the energy storage charging power at each time are used to calculate the energy saving of traction transformer loss optimization and the total energy saving. The model grid power consumption is determined based on the multi-source basic data and the traction transformer loss coefficient. An accuracy index is obtained based on the model grid power consumption and the actual grid power consumption. When the accuracy index is less than or equal to a preset accuracy threshold, an evaluation result is obtained based on the total energy saving, the photovoltaic energy saving, the energy saving from energy storage peak-valley arbitrage, the energy saving from energy storage grid loss optimization, and the energy saving from traction transformer loss optimization. Based on the proportion of energy saving in each sub-item to the total energy saving in the assessment results, the energy storage charging and discharging power correction value and the photovoltaic inverter reactive power output setting value are matched from the preset strategy library and sent to the energy storage converter and photovoltaic inverter to control the photovoltaic-storage microgrid to operate according to the corrected parameters.

[0006] By adopting the above technical solutions and utilizing multi-source data and a photovoltaic-storage collaborative model, the total energy saving can be decomposed into contributions from photovoltaics, energy storage peak-valley arbitrage, network loss optimization, and transformer loss optimization. This can reduce the aliasing interference of traction load fluctuations and seasonal differences, and relatively accurately quantify the energy-saving benefits of each link, providing a basis for railway operators to carry out refined energy-saving benefit accounting.

[0007] Optionally, the step of constructing a photovoltaic-storage synergistic energy-saving assessment model based on the multi-source basic data, the traction transformer loss coefficient, the energy storage charging and discharging efficiency, and the time period division results includes: For the missing moments in the multi-source basic data, interpolation is performed using the arithmetic mean of adjacent moments to fill in the missing moments. After removing samples that exceed a preset fluctuation threshold according to a preset dispersion criterion, the missing moments are filled in again to obtain the corrected multi-source basic data. Based on the time-series matching relationship between the photovoltaic output value, the energy storage discharge power and the traction load value at each time in the corrected multi-source basic data, dynamic calculation rules are constructed, and the dynamic calculation rules are encapsulated into the photovoltaic energy saving calculation sub-model. The photovoltaic local consumption rate at each moment output by the photovoltaic energy saving calculation sub-model is used as the criterion. Combined with the time period division results and the conversion relationship between the energy storage charging and discharging efficiency and the charging and discharging energy, a joint operation rule is constructed, and the joint operation rule is encapsulated into the energy storage energy saving calculation sub-model. The traction transformer loss coefficient is decoupled into no-load loss component and load loss component. Based on the no-load loss component and the load loss component, and combined with the equivalent load smoothing result output by the energy storage energy saving calculation sub-model, a summation operation rule is constructed, and the summation operation rule is encapsulated into the total energy saving calculation sub-model. The photovoltaic-storage synergistic energy-saving assessment model is obtained by cascading and coupling the input-output dependencies between the photovoltaic energy-saving calculation sub-model, the energy storage energy-saving calculation sub-model, and the total energy-saving calculation sub-model.

[0008] By adopting the above technical solutions, input reliability is improved through data correction. Time-series matching and local absorption rate determination are used to separate the contributions of direct photovoltaic power supply, energy storage peak-valley arbitrage, and grid loss optimization. Transformer loss decoupling and load smoothing results are combined to extract loss optimization quantities. The cascading and collaborative operation of each sub-model helps to reduce traction load fluctuation interference, distinguish various energy-saving benefits, relatively improve assessment accuracy, and provide a basis for refined benefit accounting.

[0009] Optionally, the step of calculating the photovoltaic energy saving through the photovoltaic energy saving calculation sub-model based on the photovoltaic power output value, the traction load value, and the energy storage discharge power at each time moment includes: The photovoltaic output value at each time moment is added to the energy storage discharge power at each time moment to obtain the comprehensive photovoltaic power supply at each time moment. The photovoltaic power supply at each time moment is compared with the traction load value at each time moment. When the photovoltaic power supply at each time moment is less than or equal to the traction load value at each time moment, the photovoltaic local consumption rate at each time moment is assigned a preset upper limit value of the consumption rate. When the total power supplied by the photovoltaic side at each time is greater than the traction load value at each time, the ratio of the traction load value at each time to the total power supplied by the photovoltaic side at each time is taken as the local photovoltaic consumption rate at each time. The photovoltaic energy saving at each moment is obtained by multiplying the photovoltaic output value at each moment, the photovoltaic local consumption rate at each moment, and the time step. The photovoltaic energy saving is obtained by summing the photovoltaic energy saving at each moment within the evaluation period.

[0010] By adopting the above technical solution, photovoltaic output and energy storage discharge power are integrated into a comprehensive supply. The dynamic local absorption rate is determined by comparing it with the traction load at each time step. This can relatively accurately separate the actual energy-saving contribution of photovoltaic power generation, reduce the aliasing interference of traction load fluctuations on energy-saving metering, help improve the accuracy of photovoltaic energy-saving assessment, and provide support for refined revenue accounting.

[0011] Optionally, the step of calculating the energy-saving energy storage peak-valley arbitrage energy saving and the energy-saving energy storage network loss optimization energy saving through the energy storage energy-saving calculation sub-model, based on the multi-source basic data, the energy storage charging and discharging efficiency, and the time period division results, includes: According to the time period division results, each moment in the evaluation period is divided into valley moment, flat moment and peak moment. The energy storage charging power of each moment is determined at the valley moment, the energy storage discharging power of each moment is determined at the peak moment, and the energy storage charging power and energy storage discharging power of each moment are set to zero at the flat moment. Based on the calculation relationship between the energy storage discharge power, the energy storage charging power, the energy storage charging and discharging efficiency, and the time step, the energy storage peak-valley arbitrage energy saving at each moment is obtained, and the energy storage peak-valley arbitrage energy saving at each moment within the evaluation period is accumulated in time to obtain the energy storage peak-valley arbitrage energy saving. Based on the ratio of the traction load value at each moment to the preset power supply voltage level, the equivalent current before connection at each moment is obtained; Based on the calculation relationship between the traction load value, photovoltaic output value, energy storage discharge power, energy storage charging power and the preset power supply voltage level at each time, the equivalent current after connection at each time is obtained. Based on the calculation relationship between the square of the equivalent current before access at each time point, the square of the equivalent current after access at each time point, the preset equivalent resistance of the contact network, and the time step, the energy storage network loss optimization energy saving is obtained by performing time-series accumulation within the evaluation period.

[0012] By adopting the above technical solutions, the energy storage charging and discharging behavior is divided according to peak and valley periods, and combined with the charging and discharging efficiency calculation, the peak-valley arbitrage contribution of energy storage can be separated. At the same time, by comparing the difference in the square of the equivalent current of the contact network before and after connection, the amount of network loss optimization can be quantified. This helps to distinguish the multi-dimensional energy-saving effect of energy storage from the total amount, reduce the interference of traction load fluctuations, relatively improve the accuracy of the assessment, and provide a basis for refined accounting.

[0013] Optionally, determining the energy storage charging power at each time point during the valley period and determining the energy storage discharging power at each time point during the peak period includes: At the valley time, the energy storage charging power at each time is obtained based on the preset rated energy storage charging power and combined with the preset upper limit of energy storage state of charge. At the peak time, the energy storage discharge power at each time is obtained by taking the preset rated discharge power of energy storage as a reference and combining the preset limit of energy storage state of charge. The energy storage state of charge value at each moment is obtained by performing time-series iterative calculations using the energy storage charging power, energy storage discharging power, energy storage charging and discharging efficiency, and time step.

[0014] By adopting the above technical solution, the rated power is used as the benchmark during both valley and peak periods, and upper and lower limits of the state of charge (SOC) are introduced for power correction, while the SOC value is iteratively updated. This makes the energy storage charging and discharging behavior closer to the actual operating characteristics, which helps to reduce the deviation in energy saving calculations caused by ignoring capacity limitations, thereby relatively improving the accuracy of the assessment.

[0015] Optionally, the step of calculating the total energy saving sub-model, based on the photovoltaic energy saving, the energy saving through energy storage peak-valley arbitrage, the energy saving through energy storage network loss optimization, the traction transformer loss coefficient, the traction load value at each time moment, the photovoltaic output value at each time moment, the energy storage discharge power at each time moment, and the energy storage charging power at each time moment, to obtain the traction transformer loss optimization energy saving and the total energy saving includes: Substitute the traction load values ​​at each time point into the loss calculation expression corresponding to the traction transformer loss coefficient to obtain the transformer loss values ​​before connection at each time point. Subtract the photovoltaic output value and the energy storage discharge power at each time from the traction load value at each time, and add the energy storage charging power at each time, then substitute the result into the loss calculation expression to obtain the transformer loss value after connection at each time. The difference between the transformer loss value before access at each time point and the transformer loss value after access at each time point is used as the accumulator. The accumulator is performed sequentially according to the time step within the evaluation period to obtain the energy saving of the traction transformer loss optimization. The total energy saving is obtained by summing the energy saving from photovoltaic power generation, the energy saving from energy storage peak-valley arbitrage, the energy saving from optimized energy storage network loss, and the energy saving from optimized traction transformer loss.

[0016] By adopting the above technical solution, the losses of the traction transformer before and after connection are calculated separately. After connection, the load correction caused by photovoltaic and energy storage charging and discharging is taken into account. The energy saving of transformer loss optimization is obtained by accumulating the loss difference, and then summed with the photovoltaic and energy storage components to form the total energy saving. This helps to separate the energy saving contribution of transformer operating efficiency improvement, making the energy saving structure more complete and providing a basis for refined benefit accounting.

[0017] Optionally, the step of determining the model grid power consumption based on the multi-source basic data and the traction transformer loss coefficient, obtaining an accuracy index based on the model grid power consumption and the actual grid power consumption, and when the accuracy index is less than or equal to a preset accuracy threshold, obtaining an evaluation result based on the total energy saving, the photovoltaic energy saving, the energy saving from energy storage peak-valley arbitrage, the energy saving from energy storage grid loss optimization, and the energy saving from traction transformer loss optimization, including: The photovoltaic output value and the energy storage discharge power at each time are deducted from the traction load value at each time, and the energy storage charging power at each time is added together. Then, the model grid power value at each time is calculated with the loss coefficient of the traction transformer. The model grid power value at each time is then accumulated over time within the evaluation period to obtain the model grid power consumption. The accuracy index is obtained by dividing the absolute value of the difference between the model grid power consumption and the actual grid power consumption by the actual grid power consumption. The accuracy index is compared with the preset accuracy threshold. When the accuracy index is less than or equal to the preset accuracy threshold, the total energy saving, the photovoltaic energy saving, the energy saving from energy storage peak-valley arbitrage, the energy saving from energy storage network loss optimization, and the energy saving from traction transformer loss optimization are combined according to preset evaluation dimensions to obtain the evaluation result.

[0018] By adopting the above technical solution, the net load after connection and transformer losses are used to back-calculate the grid power consumption of the model, and the accuracy index is obtained by comparing it with the actual grid power metering value. The evaluation result is only output when the accuracy meets the requirements. This verification closed loop helps to verify the rationality of the model calculation, reduces the possibility of misestimation of energy saving due to model deviation, makes the evaluation result relatively more reliable, and provides a reference for refined accounting.

[0019] A second aspect of this application provides an electronic device including a processor, a memory, a user interface, and a network interface, wherein the memory is used to store instructions, the user interface and the network interface are both used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to cause the electronic device to perform the method as described in any of the foregoing.

[0020] A third aspect of this application provides a computer-readable storage medium storing instructions that, when executed, perform the method described in any of the preceding descriptions.

[0021] A fourth aspect of this application provides a computer program product that, when run on an electronic device, causes the electronic device to perform the method as described in any of the preceding claims.

[0022] In summary, one or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: By decomposing the total energy savings into contributions from photovoltaic power generation, energy storage peak-valley arbitrage, grid loss optimization, and transformer loss optimization through a photovoltaic-storage collaborative model, various energy-saving benefits can be separated, reducing the aliasing interference from traction load fluctuations and seasonal differences, and relatively improving the accuracy of the assessment. Dynamic local absorption rate calculation, energy storage simulation under state-of-charge constraints, and transformer loss decoupling methods make the quantification of each link closer to the actual operating characteristics, which helps to reduce calculation deviations. The introduction of a closed-loop accuracy verification between the model and actual grid power metering verifies the rationality of the model, making the assessment results more credible and providing a basis for railway operators to conduct refined energy-saving benefit accounting. Attached Figure Description

[0023] Figure 1 This is a system architecture diagram of a method for assessing energy consumption after a photovoltaic-storage microgrid is connected to an electrified railway, as disclosed in an embodiment of this application. Figure 2 This is a flowchart illustrating a method for assessing energy consumption after a photovoltaic-storage microgrid is connected to an electrified railway, as disclosed in an embodiment of this application. Figure 3 This is a schematic diagram of the structure of an electronic device disclosed in an embodiment of this application.

[0024] Explanation of reference numerals in the attached figures: 100, System architecture; 101, First terminal device; 102, Second terminal device; 103, Third terminal device; 104, Network; 105, Server; 301, Processor; 302, Communication bus; 303, User interface; 304, Network interface; 305, Memory. Detailed Implementation

[0025] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0026] like Figure 1 As shown, system architecture 100 may include terminal devices 101, 102, and 103, a network 104, and a server 105. Network 104 serves as the medium for providing communication links between terminal devices 101, 102, and 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0027] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, and 103, such as model training applications, video recognition applications, web browser applications, social platform software, etc.

[0028] Terminal devices 101, 102, and 103 can be either hardware or software. When terminal devices 101, 102, and 103 are hardware, they can be various electronic devices with displays, including but not limited to smartphones, tablets, e-book readers, MP3 (Moving Picture Experts Group Audio Layer III) players, MP4 (Moving Picture Experts Group Audio Layer IV) players, laptops, and desktop computers, etc. When terminal devices 101, 102, and 103 are software, they can be installed in the aforementioned electronic devices. They can be implemented as multiple software programs or software modules (e.g., multiple software programs or software modules used to provide distributed services) or as a single software program or software module. No specific limitations are imposed here.

[0029] This embodiment discloses a method for assessing the energy consumption of a photovoltaic-storage microgrid after it is connected to an electrified railway. Figure 2 This is a flowchart illustrating a method for assessing energy savings after a photovoltaic-storage microgrid is connected to an electrified railway, as disclosed in an embodiment of this application. Figure 2 As shown, the method includes the following steps: S201. Determine the assessment boundary and set basic parameters based on the assessment boundary. The basic parameters include time step, assessment period, traction transformer loss coefficient, energy storage charging and discharging efficiency, and time period division results of railway power grid electricity price. Specifically, the evaluation boundary refers to the spatiotemporal and object scope used in this embodiment for conducting railway power grid energy flow analysis and economic evaluation. Spatially, it may include, but is not limited to, the closed system consisting of traction substations, traction networks, energy storage devices, upstream power grid access points, and the operating sections of the trains under its jurisdiction. Object-wise, it may cover equipment such as power supply side, load side, and energy storage side involved in energy exchange. There are several ways to determine the evaluation boundary: The first option is for the user to manually delineate the relevant traction substation numbers, power supply arm ranges, and energy storage device numbers on the human-computer interaction interface according to actual engineering needs; the second option is to automatically read the topology configuration file from the railway power supply dispatching system or SCADA (Supervisory Control And Data Acquisition) system and automatically generate the evaluation boundary according to preset rules (such as by line, by power supply zone, or by the area under the jurisdiction of the substation); the third option is to select the corresponding line sections as the evaluation boundary by setting the start and end mileage or latitude and longitude coordinates based on geographic information data. After determining the assessment boundaries, basic parameters are further set. These basic parameters include at least the time step Δt, the assessment period T, the traction transformer loss coefficient η_t, the energy storage charging and discharging efficiency (including charging efficiency η_c and discharging efficiency η_d), and the time period division results of the railway grid electricity price. The time step Δt represents the time granularity of data sampling and calculation during the assessment process, and can be configured according to the assessment accuracy requirements, such as 1 second, 10 seconds, 1 minute, 5 minutes, or 15 minutes. The assessment period T represents the total duration covered by the assessment, such as 1 day (24 hours), 1 week, 1 month, or 1 year, or a typical operating day. The traction transformer loss coefficient η_t is used to characterize the loss characteristics of the traction transformer during energy transfer, and its value can be obtained based on the transformer nameplate parameters, factory test reports, or historical operating data statistics. It should be noted that η_t in this application has two equivalent forms depending on the application scenario: Form 1 is the dimensionless loss rate η_t(r), which characterizes the ratio of power loss to input power, typically ranging from 0.5% to 3%, and is mainly used to calculate the equivalent energy savings from photovoltaic local consumption and energy storage charging and discharging; Form 2 is the dimensional load loss coefficient η_t(k) (unit: 1 / kW), which is numerically equal to Pk / Sn², where Pk is the rated short-circuit loss of the transformer (kW) and Sn is the rated capacity of the transformer (kVA), and is mainly used to calculate the transformer copper loss value time-by-time in the form of η_t(k)·P². The two forms are used in this application according to the specific context and are not intended to cause ambiguity.Unless otherwise specified, η_t appearing in this application shall be understood according to the expression corresponding to the context. The energy storage charge and discharge efficiency is used to characterize the efficiency of electrical energy conversion during the charging and discharging process of the energy storage device. It can be determined according to the energy storage type (such as lithium-ion battery, supercapacitor, flywheel energy storage, etc.) and the technical parameters provided by its manufacturer. For example, the typical range of η_c and η_d for lithium-ion batteries is 0.90~0.98. The time period division results of railway power grid electricity price are used to reflect the purchase and sale prices of electricity in different time periods. The division method can adopt a three-stage peak-valley-flat period (e.g., peak period 8:00-11:00 and 18:00-21:00, flat period 7:00-8:00 and 11:00-18:00, valley period 21:00-7:00 the next day), or a four-stage peak-peak-flat-valley period, or be configured according to the time-of-use electricity price policy actually published by the local power supply company. The specific values ​​of the above basic parameters can be flexibly configured according to the actual application scenario, regional policies and equipment characteristics. This embodiment does not limit this.

[0030] S202. Collect multi-source basic data within the assessment period according to the time step. The multi-source basic data includes the traction load value, photovoltaic output value, energy storage charging power, energy storage discharging power and grid power metering value at each time. The grid power metering value at each time within the assessment period is accumulated to obtain the actual grid power consumption. Specifically, multi-source basic data refers to a set of operational data from different measurement objects and different types of physical quantities. In this embodiment, it includes at least the traction load value P_L(t) at each time, the photovoltaic output value P_pv(t) at each time, the energy storage charging power P_ch(t) at each time, the energy storage discharging power P_dis(t) at each time, and the grid power metering value P_g(t) at each time, where t represents the sampling time after discretization according to the time step Δt. There are several ways to collect the above data: The first approach is to directly collect power and energy signals using power quality analyzers, smart meters, or multi-functional measurement and control devices deployed at the incoming line side of traction substations, photovoltaic grid connection points, AC side of energy storage converters, and metering points, and upload them to the data acquisition server via communication protocols such as IEC 61850, Modbus TCP, DL / T 645, or DNP 3.0; The second approach is to periodically extract the corresponding operating data from the historical database of the railway power supply SCADA system, energy management system, or photovoltaic / energy storage local monitoring system, keeping the collection frequency consistent with the time step Δt; The third approach is to subscribe to the real-time data stream of the corresponding measurement point on the data bus or message middleware (such as Kafka, MQTT, etc.), downsample or aggregate it according to Δt, and then store it in a time series database (such as InfluxDB, TDengine, etc.). During data acquisition, necessary preprocessing can be further performed, such as, but not limited to, missing value imputation (which can be done using forward imputation, linear interpolation, or imputation based on the average of historical time periods), outlier removal (which can be done using the 3σ criterion, box plot method, or limit judgment based on the rated capacity of the equipment), and clock synchronization correction (which can be based on NTP or GPS time synchronization) to ensure the consistency and comparability of multi-source data in the time dimension. After data acquisition and preprocessing are completed, the grid power consumption P_g(t) at each time point within the evaluation period T is summed over a time step Δt, i.e., E_g = ΣP_g(t)·Δt, t∈[0,T], thereby obtaining the actual grid power consumption E_g; the above sampling time step, evaluation period length, and communication protocol and deployment location of each measuring point can be flexibly configured according to actual engineering conditions, and this embodiment does not limit this.

[0031] S203. Based on multi-source basic data, traction transformer loss coefficient, energy storage charging and discharging efficiency and time period division results, a photovoltaic-storage synergistic energy saving assessment model is constructed. The photovoltaic-storage synergistic energy saving assessment model includes a photovoltaic energy saving calculation sub-model, an energy storage energy saving calculation sub-model and a total energy saving calculation sub-model. Specifically, the photovoltaic-storage synergistic energy-saving assessment model refers to a set of mathematical expressions used to quantify the energy-saving benefits generated by the synergistic operation of photovoltaic power generation and energy storage devices. In this embodiment, it includes at least three parts: a photovoltaic energy-saving calculation sub-model, an energy storage energy-saving calculation sub-model, and a total energy-saving calculation sub-model. The photovoltaic energy-saving calculation sub-model is used to characterize the grid power consumption reduction caused by photovoltaic output replacing the power supply from the upstream grid. One specific implementation can be E_PV=ΣP_pv(t)·(1+η_t)·Δt, t∈[0,T], where the (1+η_t) term is used to calculate the traction transformer losses avoided due to local photovoltaic consumption. As an alternative, the losses can also be restored using the form E_PV=ΣP_pv(t)·Δt / (1-η_t), or when the photovoltaic is not directly connected to the traction side. Instead, this loss conversion term is omitted in scenarios connected to medium-voltage busbars; the energy storage energy saving calculation sub-model is used to characterize the energy saving and economic benefits brought about by energy storage devices through valley charging and peak discharging or absorption of regenerative braking energy, etc., and one specific implementation can be E_ES=Σ[P_dis(t)·η_d·(1+η_t)−P_ch(t) / η_c·(1+η_t)]·Δt, t∈[0,T], that is, the net output power after deducting the charging power from the discharge power is used as the measure of energy storage energy saving, E_ES in In subsequent embodiments, the energy-saving amounts are further subdivided into energy storage peak-valley arbitrage energy saving (Warb) and energy storage network loss optimization energy saving (W_loss), i.e., E_ES = Warb + W_loss; where Warb corresponds to the electricity price difference conversion part, and W_loss corresponds to the network loss reduction part brought about by load smoothing. As an alternative, the charging and discharging amounts can also be weighted and summed according to peak, flat, and valley periods, i.e., E_ES = Σ(k_i_dis·E_dis_i−k_i_ch·E_ch_i), i∈{p, f In the formula, E_ES is the net energy saving of the energy storage device (kWh); i is the time period index, i∈{p,f,v} corresponds to the peak, flat and valley periods respectively; E_dis_i and E_ch_i are the cumulative discharge and cumulative charging power of the energy storage device in the i-th time period (kWh); k_i_dis and k_i_ch are the weighting coefficients of the discharge side and the charging side in the i-th time period, respectively, to reflect the differentiated contribution of time-of-use pricing or regenerative braking energy recovery to energy saving benefits; Σ represents summation over index i. The total energy saving calculation sub-model is used to comprehensively characterize the overall energy saving effect under the synergistic effect of photovoltaic and energy storage. One specific implementation method is E_total=E_PV+E_ES. As an alternative, a synergistic correction coefficient λ can also be introduced, such that E_total=λ·(E_PV+E_ES), where λ is used to reflect the additional gain or loss of photovoltaic and energy storage joint scheduling relative to independent operation. Its value can be determined by regression calibration based on historical operation data or based on simulation results, and the typical value range is, for example, 0.95~1.05.The specific construction form of the photovoltaic-storage synergistic energy-saving assessment model can be flexibly selected according to the access location, topology and operation strategy of photovoltaic and energy storage in actual projects. This embodiment does not limit this.

[0032] Optionally, based on multi-source basic data, traction transformer loss coefficient, energy storage charging and discharging efficiency, and time period division results, a photovoltaic-energy storage synergistic energy-saving assessment model is constructed. This includes: for missing moments in the multi-source basic data, interpolation is performed using the arithmetic mean of adjacent moments to complete the data, and samples exceeding a preset fluctuation threshold are removed according to a preset dispersion criterion before being completed again to obtain corrected multi-source basic data; based on the time-series matching relationship between photovoltaic power output, energy storage discharge power, and traction load values ​​at each moment in the corrected multi-source basic data, dynamic calculation rules are constructed, and these rules are encapsulated into a photovoltaic energy-saving calculation sub-model; the photovoltaic energy-saving calculation sub-model outputs the energy-saving data at each moment... The local photovoltaic (PV) grid connection rate is used as the criterion. Combining the time period division results and the conversion relationship between energy storage charging and discharging efficiency and charging and discharging energy, a joint calculation rule is constructed and encapsulated into an energy storage energy saving calculation sub-model. The traction transformer loss coefficient is decoupled into no-load loss component and load loss component. Based on the no-load loss component and load loss component, and combined with the equivalent load smoothing result output by the energy storage energy saving calculation sub-model, a summation calculation rule is constructed and encapsulated into a total energy saving calculation sub-model. According to the input-output dependency relationship between the PV energy saving calculation sub-model, the energy storage energy saving calculation sub-model, and the total energy saving calculation sub-model, a cascaded coupling is performed to obtain the PV-storage synergistic energy saving evaluation model.

[0033] Specifically, for missing moments in the multi-source basic data, interpolation is performed using the arithmetic mean of adjacent moments to complete the data. Then, samples exceeding a preset fluctuation threshold are removed according to a preset dispersion criterion before being filled again, resulting in corrected multi-source basic data. Specifically, for any missing data point X at sampling time t_k, interpolation can be performed using X(t_k) = [X(t_{k-1}) + X(t_{k+1})] / 2. Alternatively, a weighted arithmetic mean of n adjacent moments can be used, i.e., X(t_k) = Σw_i·X(t_{k±i}), where the weight w_i can be allocated according to the reciprocal of the time distance. The preset dispersion criterion can be implemented in several ways: the first approach is based on the 3σ criterion, which identifies samples that deviate from the sliding window mean by more than 3 times the standard deviation as outliers; the second approach is based on the box plot method, which identifies samples that exceed the range of [Q1−1.5·IQR, Q3+1.5·IQR] as outliers; after removing outliers, the above interpolation method is used again to complete the data, thereby obtaining corrected multi-source basic data with complete time series and numerical smoothness.

[0034] Furthermore, based on the time-series matching relationship between photovoltaic power output, energy storage discharge power, and traction load values ​​at each time in the corrected multi-source basic data, dynamic calculation rules are constructed, and these rules are encapsulated into a photovoltaic energy-saving calculation sub-model. The time-series matching relationship refers to the instantaneous supply and demand correspondence between photovoltaic power output P_pv(t), energy storage discharge power P_dis(t), and traction load P_L(t) at the same sampling time t. One specific scheme for constructing dynamic calculation rules is as follows: At each time t, first determine the relationship between P_pv(t) and P_L(t). When P_pv(t) ≤ P_L(t), all photovoltaic output is consumed locally, and the photovoltaic energy saving contribution is P_pv(t)·(1+η_t)·Δt. When P_pv(t) > P_L(t), the excess is either absorbed through energy storage or fed back to the grid. The energy saving contribution of the locally consumed portion is calculated as P_L(t)·(1+η_t)·Δt, while the fed-back portion is calculated separately as P_pv(t)·Δt − P_L(t)·(1+η_t)·Δt. Encapsulating the above dynamic calculation logic at the code level in the form of functions, classes, or independent calculation modules yields the photovoltaic energy saving calculation sub-model.

[0035] Furthermore, using the local photovoltaic absorption rate at each moment output by the photovoltaic energy-saving calculation sub-model as the criterion, and combining the time period division results and the conversion relationship between energy storage charging and discharging efficiency and charging and discharging energy, a joint calculation rule is constructed, and the joint calculation rule is encapsulated into an energy storage energy-saving calculation sub-model. Specifically, the local absorption rate α_PV is used as one of the inputs for determining the energy storage operating condition: when α_PV<1 (i.e., there is surplus photovoltaic output) and t is in the valley period, the energy storage prioritizes charging, and its energy-saving contribution is calculated as the charging amount P_ch(t)·Δt / η_c after conversion; when α_PV=1 and t is in the peak period, the energy storage prioritizes discharging, and its energy-saving contribution is calculated as P_dis(t)·η_d·(1+η_t)·Δt. Based on the time period division results, the evaluation period T is divided into peak, flat, and valley time periods T_p, T_f, and T_v. Then, a specific way to calculate the energy saving of energy storage can be: E_ES=Σ_{t∈T_p}P_dis(t)·η_d·(1+η_t)·Δt−Σ_{t∈T_v}P_ch(t) / η_c·(1+η_t)·Δt+ΔE_r, where ΔE_r is an additional term for regenerative braking energy recovery.

[0036] Furthermore, the traction transformer loss coefficient is decoupled into no-load loss components and load loss components. Based on these components and combined with the equivalent load mitigation results output by the energy storage energy-saving calculation sub-model, a summation operation rule is constructed and encapsulated into a total energy-saving calculation sub-model. Specifically, the traction transformer loss coefficient η_t is decoupled as η_t = η_0 + η_k · β(t)², where η_0 is the no-load loss component, whose value is only related to the transformer core characteristics and can be taken as, for example, 0.1%~0.3% based on the nameplate parameters; η_k is the load loss coefficient; β(t) = S(t) / S_N is the load rate; S(t) is the actual apparent power at time t; and S_N is the rated capacity. The equivalent load smoothing result refers to the equivalent load curve P_eq(t) = P_L(t) − P_pv(t) − P_dis(t) + P_ch(t) borne by the traction transformer after the energy storage charging and discharging operation. Its smoothing effect is reflected by the decrease in the fluctuation amplitude of the load rate β(t). A specific way to calculate the total energy saving is E_total = E_PV + E_ES + ΔE_loss, where ΔE_loss = Σ[η_k·(β_0(t)²−β_eq(t)²)·S_N]·Δt is the load loss reduced due to load smoothing, and β_0(t) and β_eq(t) are the load rates before and after smoothing, respectively.

[0037] Furthermore, based on the input-output dependencies between the photovoltaic energy saving calculation sub-model, the energy storage energy saving calculation sub-model, and the total energy saving calculation sub-model, a cascaded coupling is performed to obtain the photovoltaic-storage synergistic energy saving assessment model. Specifically, the cascaded coupling method is as follows: the photovoltaic energy saving calculation sub-model takes the corrected multi-source basic data as input and outputs the photovoltaic local absorption rate α_PV and photovoltaic energy saving E_PV at each time point; the energy storage energy saving calculation sub-model takes α_PV, the time period division results, and η_c and η_d as input and outputs the energy storage energy saving E_ES and the equivalent load curve P_eq(t); the total energy saving calculation sub-model takes E_PV, E_ES, and P_eq(t) as input and outputs the total energy saving E_total. There are several ways to implement cascaded coupling: the first approach is to chain the three sub-models together using a function call chain, where the return value of the previous sub-model is used as the input parameter of the next sub-model; the second approach is to use a directed acyclic graph scheduling framework (such as Airflow, Perfect, etc.) to orchestrate the tasks of the three sub-models; after the cascaded coupling is completed, the resulting photovoltaic-storage collaborative energy-saving assessment model can provide a unified input and output interface for subsequent economic evaluation and other steps.

[0038] S204. The photovoltaic energy saving is calculated using the photovoltaic energy saving calculation sub-model based on the photovoltaic output value, traction load value, and energy storage discharge power at each time. Specifically, the corrected photovoltaic output value P_pv(t), traction load value P_L(t), and energy storage discharge power P_dis(t) at each time are sequentially input into the constructed photovoltaic energy saving calculation sub-model according to the time step Δt. At each sampling time t, point-by-point judgment and summation are performed according to the aforementioned time sequence matching relationship to obtain the photovoltaic energy saving E_PV within the evaluation period T. One specific method for calculating photovoltaic energy saving is as follows: First, at each time t, calculate the upper limit of photovoltaic power that can be locally absorbed, P_cap(t) = max[P_L(t) − P_dis(t), 0], which is the power gap remaining after deducting energy storage discharge from the traction load. Then, take P_use(t) = min[P_pv(t), P_cap(t)] as the photovoltaic power absorbed locally, and its corresponding energy saving contribution is P_use(t)·(1+η_t)·Δt. The excess P_pv(t) − P_use(t) (if it exists) is included in P_res(t)·Δt according to the path that does not pass through the traction transformer. Finally, sum over the entire evaluation period to obtain E_PV = Σ[P_use(t)·(1+η_t)]. )+P_res(t)]·Δt, t∈[0,T]; As an alternative, the local absorption rate α_PV=P_use(t) / P_pv(t) can also be used to represent it, that is, E_PV=ΣP_pv(t)·[α_PV·(1+η_t)+(1−α_PV)]·Δt, or directly simplified to E_PV=ΣP_pv(t)·(1+η_t)·Δt in the scenario where the photovoltaic output is always less than the net load; Another option is to introduce the photovoltaic module efficiency correction factor k_PV (for example, take 0.95~0.99 to reflect module aging, shading and inverter loss) and then perform integration and summation, that is, E_PV=Σk_PV·P_pv(t)·(1+η_t)·Δt. The above calculation process can be implemented in various ways, such as loop traversal, vectorized array operation (e.g., based on numerical computing libraries such as NumPy and Pandas), or database-side SQL aggregation. This embodiment does not limit this. The calculated photovoltaic energy saving E_PV serves as one of the input components for the total energy saving calculation. On the other hand, its time-series intermediate quantity α_PV and P_use(t) can serve as the basis for the subsequent energy storage energy saving calculation sub-model, thereby realizing the organic connection between photovoltaic and energy storage calculation links.

[0039] Optionally, the photovoltaic energy saving calculation sub-model calculates the photovoltaic energy saving based on the photovoltaic output value, traction load value, and energy storage discharge power at each time point. This includes: adding the photovoltaic output value and energy storage discharge power at each time point to obtain the comprehensive photovoltaic power supply at each time point; comparing the comprehensive photovoltaic power supply at each time point with the traction load value at each time point; when the comprehensive photovoltaic power supply at each time point is less than or equal to the traction load value, assigning the local photovoltaic absorption rate at each time point to a preset upper limit value; when the comprehensive photovoltaic power supply at each time point is greater than the traction load value, using the ratio of the traction load value to the comprehensive photovoltaic power supply at each time point as the local photovoltaic absorption rate at each time point; obtaining the photovoltaic energy saving at each time point by multiplying the photovoltaic output value, the local photovoltaic absorption rate at each time point, and the time step; and accumulating the photovoltaic energy saving at each time point within the evaluation period to obtain the total photovoltaic energy saving.

[0040] Specifically, the photovoltaic output value P_pv(t) at each time moment is added to the energy storage discharge power P_dis(t) at each time moment to obtain the comprehensive photovoltaic power supply at each time moment P_sup(t) = P_pv(t) + P_dis(t). The comprehensive photovoltaic power supply refers to the total active power provided by photovoltaics and energy storage to the traction load side at the same sampling time t. Its physical meaning is the amount of local clean energy supply that can be used to replace the power supply from the upper-level grid. This addition operation can be implemented in various ways, such as point-by-point cyclic accumulation, array vectorization operation, or database field addition. This embodiment does not limit this method.

[0041] Furthermore, the comprehensive photovoltaic power supply P_sup(t) at each moment is compared with the traction load value P_L(t) at each moment, and the local photovoltaic absorption rate α_PV at each moment is assigned a value based on the comparison result. Specifically, when P_sup(t)≤P_L(t), it indicates that the total supply of photovoltaic and energy storage is insufficient to cover the current traction load demand, and the photovoltaic output can be completely absorbed locally by the traction load. At this time, α_PV is assigned the preset absorption rate upper limit value α_max. One specific value of α_max is 1.0, which means that the photovoltaic output is considered to be 100% absorbed locally. As an alternative, considering the line loss and a small amount of backfeed from the grid connection interface, α_max can also be taken as a value slightly less than 1 (e.g., 0.98~0.995). Its specific value can be configured according to the actual application requirements and is not limited here. When P_sup(t) > P_L(t), it indicates that the total supply of photovoltaic and energy storage has exceeded the current traction load demand, and some power cannot be absorbed locally. At this time, the local absorption rate is assigned a value according to α_PV = P_L(t) / P_sup(t). This ratio physically represents the proportion actually absorbed by the traction load in the overall supply. There are several ways to implement the above comparison and determination: the first approach is to use if-else conditional branches to determine the value point by point; the second approach is to use vectorized piecewise functions (such as numpy.where or pandas.mask) for batch calculation.

[0042] Furthermore, based on the product of the photovoltaic output value P_pv(t) at each moment, the local photovoltaic absorption rate α_PV at each moment, and the time step Δt, the photovoltaic energy saving ΔE_PV(t) = P_pv(t)·α_PV·Δt at each moment is obtained. The time step Δt is consistent with the sampling interval of data acquisition in step S202, for example, it can be 1s, 1min, 5min or 15min. Its specific value can be configured according to the balance between evaluation accuracy and computational load, and is not limited here. As an alternative, to further account for the traction transformer loss avoided by local photovoltaic absorption, the product form can also be extended to ΔE_PV(t) = P_pv(t)·α_PV·(1+η_t)·Δt, where η_t is the traction transformer loss coefficient.

[0043] Furthermore, the photovoltaic energy saving ΔE_PV(t) at each moment within the evaluation period T is accumulated over time to obtain the photovoltaic energy saving E_PV = Σ_{t=0}^{T}ΔE_PV(t). One specific implementation scheme for time-series accumulation is to perform discrete summation on the set of sampling times; as an alternative, numerical integration methods such as trapezoidal integral and Simpson integral can be used to improve the accumulation accuracy when the sampling interval is large; another optional scheme is to first accumulate the photovoltaic energy saving E_PV_p, E_PV_f, and E_PV_v separately according to peak, flat, and valley periods, and then sum the three to obtain the total photovoltaic energy saving E_PV, thereby providing intermediate results for subsequent economic evaluation.

[0044] S205. Through the energy storage energy saving calculation sub-model, based on multi-source basic data, energy storage charging and discharging efficiency and time period division results, the energy saving of energy storage peak-valley arbitrage and the energy saving of energy storage network loss optimization are calculated. Specifically, the collected and corrected energy storage charging power P_ch(t), energy storage discharging power P_dis(t), traction load value P_L(t), and photovoltaic output value P_pv(t) at each time moment, together with the energy storage charging efficiency η_c and discharging efficiency η_d set in step S201, and the output peak, flat, and valley time sets T_p, T_f, and T_v, are input into the energy storage energy saving calculation sub-model constructed in S203, and point-by-point judgment and item-by-item accumulation are performed according to the time step Δt within the evaluation period T. Warb, the energy saving through peak-valley arbitrage in energy storage, refers to the equivalent energy saving calculated by the electricity price difference achieved through peak-valley charging and peak-valley discharging. One specific calculation method is as follows: First, during the peak period t∈T_p, the discharge capacity E_dis_p is obtained by accumulating P_dis(t)·η_d·Δt; then, during the valley period t∈T_v, the charging capacity E_ch_v is obtained by accumulating P_ch(t) / η_c·Δt. Warb = E_dis_p−E_ch_v is then used to characterize the arbitrage energy saving. As an alternative, it is also... Weighting coefficients k_p and k_v for peak and valley periods can be introduced separately (e.g., set according to the peak-valley electricity price ratio), i.e., Warb=k_p·E_dis_p−k_v·E_ch_v, or a weighting term for the normal period can be further introduced to support the three-stage electricity price structure; the energy saving W_loss of energy storage network loss optimization refers to the reduction of network loss due to the smoothing effect of energy storage charging and discharging on the equivalent load of traction transformers. One specific calculation method is: first construct the original equivalent load P_0(t)=P_L(t) before energy storage operation. −P_pv(t) and the equivalent load after energy storage action P_eq(t) = P_L(t) − P_pv(t) − P_dis(t) + P_ch(t) are then accumulated based on the load loss characteristics of the traction transformer using W_loss = Ση_k·[β_0(t)²−β_eq(t)²]·S_N·Δt, where β_0(t) = P_0(t) / S_N and β_eq(t) = P_eq(t) / S_N are the load rates before and after smoothing, respectively, and η_k is the load. The loss coefficient and S_N represent the rated capacity of the transformer. The above calculation process can be implemented in various ways, such as loop traversal, vectorized array operation (e.g., based on numerical computing libraries such as NumPy and Pandas), or database-side SQL grouping and aggregation. The obtained Warb and W_loss will be used as input components of the subsequent total energy calculation sub-model. At the same time, the intermediate quantity P_eq(t) will also be passed to the total energy calculation sub-model as the equivalent load smoothing result to support further refined calculation of transformer losses.

[0045] Optionally, using an energy storage energy-saving calculation sub-model, based on multi-source basic data, energy storage charging and discharging efficiency, and time period division results, the energy-saving energy from peak-valley arbitrage and the energy-saving energy from network loss optimization are calculated. This includes: dividing each moment within the evaluation period into valley, flat, and peak times according to the time period division results; determining the energy storage charging power at each moment during the valley time; determining the energy storage discharging power at each moment during the peak time; and setting both the energy storage charging and discharging power to zero at each moment during the flat time; and obtaining the energy-saving energy from peak-valley arbitrage at each moment based on the operational relationships between the energy storage discharging power, the energy storage charging power, the energy storage charging and discharging efficiency, and the time step. The energy-saving peak-valley arbitrage at each moment within the assessment period is accumulated over time to obtain the energy-saving peak-valley arbitrage. Based on the ratio of the traction load value to the preset power supply voltage level at each moment, the equivalent current before connection at each moment is obtained. Based on the calculation relationship between the traction load value, photovoltaic output value, energy storage discharge power, energy storage charging power, and preset power supply voltage level at each moment, the equivalent current after connection at each moment is obtained. Based on the calculation relationship between the square of the equivalent current before connection at each moment, the square of the equivalent current after connection at each moment, the preset contact network equivalent resistance, and the time step, the energy-saving network loss optimization is accumulated over the assessment period to obtain the energy-saving network loss optimization.

[0046] Specifically, according to the time period division results output by S203, each sampling time t within the evaluation period T is divided into a valley time set T_v, a flat time set T_f, and a peak time set T_p. At a valley time t∈T_v, the energy storage charging power P_ch(t) is determined based on the energy storage SOC and charging power data collected by S202, and P_dis(t) is set to zero. At a peak time t∈T_p, the energy storage discharging power P_dis(t) is determined based on the energy storage SOC and discharging power data, and P_ch(t) is set to zero. At a flat time t∈T_f, both P_ch(t) and P_dis(t) are set to zero, meaning that energy storage does not participate in charging or discharging during flat times.

[0047] Furthermore, based on the operational relationships between the energy storage discharge power P_dis(t), the energy storage charging power P_ch(t), the energy storage charging and discharging efficiency (η_c, η_d), and the time step Δt at each moment, the energy storage peak-valley arbitrage saving energy ΔWarb(t) at each moment is obtained. One specific calculation scheme is as follows: during peak hours, the equivalent energy-saving contribution on the discharge side is calculated as ΔWarb(t) = P_dis(t)·η_d·Δt; during off-peak hours, the equivalent power consumption deduction on the charging side is calculated as ΔWarb(t) = −P_ch(t) / η_c·Δt; and during flat hours, ΔWarb(t) = 0. As an alternative, peak-valley electricity price weighting coefficients k_p and k_v can be introduced on both sides respectively (for example, set according to the peak-valley electricity price ratio), i.e., ΔWarb(t) = k_p·P_dis(t)·η_d·Δt−k_v·P_ch(t) / η_c·Δt, to support arbitrage assessment based on electricity tariff calculation. Then, the energy saved by arbitrage at each moment within the evaluation period T is accumulated over time to obtain the energy saved by peak-valley arbitrage Warb=Σ_{t=0}^{T}ΔWarb(t). The accumulation method can be implemented by discrete summation, trapezoidal integration, or accumulation and summation in different time periods. This embodiment does not limit the implementation method.

[0048] Further, the calculation of energy-saving amount for energy storage network loss optimization is performed. Based on the ratio of the traction load value P_L(t) at each moment to the preset power supply voltage level U (e.g., DC 1500V or 750V for urban rail transit, the specific value can be configured according to the line system), the equivalent current I_0(t) = P_L(t) / U before connection is obtained. The equivalent current before connection represents the equivalent DC current flowing through the contact network when neither photovoltaic nor energy storage is connected. As an alternative, when the power supply system adopts AC system, it can also be converted according to I_0(t) = P_L(t) / (sqrt(3)·U·cosφ), where cosφ is the preset power factor, the specific value of which can be configured according to the actual application requirements, and is not limited here.

[0049] Furthermore, based on the operational relationships between the traction load value P_L(t), the photovoltaic output value P_pv(t), the energy storage discharge power P_dis(t), the energy storage charging power P_ch(t), and the preset power supply voltage level U at each moment, the equivalent current I_1(t) after connection at each moment is obtained as follows: I_1(t) = [P_L(t) − P_pv(t) − P_dis(t) + P_ch(t)] / U. The equivalent current after connection represents the equivalent DC current flowing through the contact network after the photovoltaic and energy storage systems are jointly connected and participate in the operation. Since the photovoltaic output and energy storage discharge offset part of the traction load demand, while energy storage charging increases the load demand, the algebraic calculation of the above four power values ​​can reflect the net impact of local clean energy connection on the contact network power flow. When the power supply system adopts an AC system, the denominator can be adjusted accordingly using the same conversion formula as described above.

[0050] Based on the calculation relationship between the square of the equivalent current before connection I_0²(t) and the square of the equivalent current after connection I_1²(t) at each time, the preset equivalent resistance of the contact network R_eq, and the time step Δt, time-series accumulation is performed within the evaluation period T to obtain the energy saving of energy storage network loss optimization. The equivalent resistance R_eq of the contact network can be calculated using the formula R_eq=ρ·L / A based on parameters such as line length, conductor cross-sectional area, material, and operating temperature. Here, ρ represents the resistivity of the conductor material at the operating temperature, L represents the length of the contact network line, and A represents the cross-sectional area of ​​the conductor. The resistivity ρ can be further corrected based on material properties and a temperature correction factor using the formula ρ(T)=ρ0·[1+α_R·(T−T0)], where ρ0 is the resistivity at the reference temperature T0, α_R is the temperature coefficient of resistance, and T is the actual operating temperature.

[0051] Optionally, the energy storage charging power at each time point is determined during off-peak hours, and the energy storage discharging power at each time point is determined during peak hours. This includes: during off-peak hours, obtaining the energy storage charging power at each time point based on a preset rated energy storage charging power and constrained by a preset upper limit of the energy storage state of charge; during peak hours, obtaining the energy storage discharging power at each time point based on a preset rated energy storage discharging power and constrained by a preset upper limit of the energy storage state of charge; and obtaining the energy storage state of charge value at each time point through time-series iterative calculation using the energy storage charging power at each time point, the energy storage discharging power at each time point, the energy storage charging and discharging efficiency, and the time step.

[0052] During off-peak hours, the energy storage charging power P_ch(t) is determined at each time point based on the preset rated energy storage charging power Pchn and in combination with the preset upper limit of energy storage state of charge SOCmax. In one optional implementation, it can be calculated using the formula P_ch(t) = min{Pchn, [SOCmax−SOC(t−Δt)]·EN / (ηch·Δt)}, where P_ch(t) is the energy storage charging power at time t, in kW; Pchn is the preset rated energy storage charging power, in kW; SOCmax is the preset upper limit of the energy storage state of charge, ranging from 0 to 1, for example, 0.9 or 0.95; SOC(t−Δt) is the energy storage state of charge value at the previous time, dimensionless, ranging from 0 to 1; EN is the rated capacity of the energy storage system, in kWh; ηch is the energy storage charging efficiency, dimensionless, for example, 0.90 to 0.98; and Δt is the time step, in hours. The physical meaning of the above formula is: under normal circumstances, constant power charging is performed at the rated power. When the remaining rechargeable capacity is insufficient to support the rated power charging for a period of Δt, the derating is automatically reduced to prevent the SOC from exceeding the upper limit.

[0053] At peak times, the energy storage discharge power P_dis(t) is determined at each time point, based on the preset rated discharge power Pdisn and constrained by the preset state of charge (SOC) limit. In one optional implementation, it can be calculated using the formula P_dis(t) = min{Pdisn, [SOC(t−Δt)−SOCmin]·EN·ηdis / Δt}, where P_dis(t) is the energy storage discharge power at time t (kW); Pdisn is the preset rated discharge power (kW); SOCmin is the preset state of charge limit, ranging from 0 to 1 (e.g., 0.1 or 0.2); ηdis is the energy storage discharge efficiency, dimensionless (e.g., 0.90 to 0.98); and the remaining parameters have the same meaning. The physical meaning of this formula is: under normal conditions, the battery discharges at a constant rated power. When the remaining discharge capacity is insufficient to support the rated power discharge for a period Δt, derating is performed to prevent the SOC from exceeding the lower limit and damaging battery life.

[0054] It should be noted that the specific values ​​of the parameters Pchn, Pdisn, SOCmax, SOCmin, ηch, ηdis, EN, and Δt mentioned above can be configured according to the selection specifications of the energy storage device and the actual engineering operation requirements, and are not limited here. As alternative implementation schemes, the following methods can also be adopted: The first scheme is to directly set P_ch(t) = Pchn or P_dis(t) = Pdisn during the valley or peak periods, and operate at constant power at rated power, cutting off charging or discharging only when the upper or lower limit of SOC is triggered; the second scheme is to evenly distribute the power according to its duration and available capacity during the valley or peak periods to achieve equal power balanced charging and discharging; the third scheme is to introduce a power-SOC segmented droop characteristic curve, so that the power linearly decays to zero when the SOC approaches the upper or lower limit, in order to further extend the battery life.

[0055] After obtaining the energy storage charging power P_ch(t) and energy storage discharging power P_dis(t) at each time point, further time-series iterative calculations are performed using the energy storage charging power, energy storage discharging power, energy storage charging and discharging efficiency, and time step to obtain the energy storage state of charge value SOC(t) at each time point. In one optional implementation, the charging period is iterated according to the formula SOC(t) = SOC(t−Δt) + P_ch(t)·ηch·Δt / EN, the discharging period is iterated according to the formula SOC(t) = SOC(t−Δt) − P_dis(t)·Δt / (ηdis·EN), and other periods where neither charging nor discharging occurs are kept unchanged with SOC(t) = SOC(t−Δt). Wherein, SOC(t) is the energy storage state of charge value at time t, dimensionless, ranging from 0 to 1; SOC(t−Δt) is the energy storage state of charge value at the previous time; P_ch(t) and P_dis(t) are the energy storage charging power and discharging power at time t, respectively, in kW; ηch and ηdis are the energy storage charging efficiency and discharging efficiency, respectively, both dimensionless; EN is the rated capacity of the energy storage system, in kWh; Δt is the time step, in h. The initial value SOC(0) for iterative calculation can be preset, for example, SOC(0) = SOCmin, or the initial state of charge value specified by the user, which is not limited here.

[0056] As alternative schemes for the above-mentioned time-series iterative calculation, the first scheme adopts a unified expression SOC(t) = SOC(t−Δt) + [P_ch(t)·ηch−P_dis(t) / ηdis]·Δt / EN, which is compatible with the two operating conditions by setting P_dis(t) = 0 and P_ch(t) = 0 during the charging and discharging periods respectively, which is convenient for programmatic implementation; the second scheme combines the charging efficiency and discharging efficiency and converts them into a comprehensive round-trip efficiency ηrt = ηch·ηdis before performing equivalent iteration; the third scheme introduces a correction term for the battery self-discharge rate γ during the iteration process, that is, SOC(t) = SOC(t−Δt)·(1−γ·Δt) + ΔSOC(t), to more accurately reflect the energy decay over a long time scale, where γ is the self-discharge rate of the energy storage battery, and its specific value can be configured according to the measured parameters of the energy storage battery, which is not limited here.

[0057] S206. Using the total energy saving calculation sub-model, the energy saving of photovoltaic, energy saving of energy storage peak-valley arbitrage, energy saving of energy storage network loss optimization, traction transformer loss coefficient, traction load value at each time, photovoltaic output value at each time, energy storage discharge power at each time, and energy storage charging power at each time are calculated to obtain the energy saving of traction transformer loss optimization and the total energy saving. Using the total energy saving calculation sub-model, based on the obtained photovoltaic energy saving E_PV, energy storage peak-valley arbitrage energy saving Warb, energy storage network loss optimization energy saving W_loss, traction transformer loss coefficient η_t (whose physical meaning is the ratio coefficient between the copper loss of the traction transformer and the square of the active power of the load it bears, which can be obtained by fitting the traction transformer nameplate parameters or historical operating data, for example, η_t=k / Un², where k is the transformer short-circuit loss coefficient and Un is the transformer rated voltage), traction load value P_L(t) at each time, photovoltaic output value P_pv(t) at each time, energy storage discharge power P_dis(t) at each time, and energy storage charging power P_ch(t) at each time, the traction transformer loss optimization energy saving Wtr and the total energy saving E_total are calculated. In one optional implementation, the energy saving from optimized traction transformer losses can be calculated using the formula Wtr=Σ{η_t·[P_L(t)]²−η_t·[P_L(t)−P_pv(t)−P_dis(t)+P_ch(t)]²}·Δt. This means that the copper loss of the traction transformer before it is connected to the photovoltaic energy storage system is taken as the benchmark, and the copper loss generated by the actual equivalent load borne by the traction transformer after connecting to the photovoltaic energy storage system is subtracted. The difference between the two is summed over the entire statistical period at a time step Δt to obtain Wtr, with the unit being kWh. The summation range in the formula can be a statistical period of 24 hours a day, a month, or a year, etc., without any specific limitation. The total energy savings are calculated using the formula E_total = E_PV + Warb + W_loss + Wtr, where E_PV is the energy savings from photovoltaic power generation itself, Warb is the equivalent energy savings from energy storage through arbitrage using peak-valley electricity price differences, W_loss is the energy savings from energy storage by optimizing power flow distribution and reducing feeder network losses, Wtr is the energy savings from transformer copper losses by reducing load fluctuations borne by the traction transformer, and E_total is the sum of the above four parts, all in kWh. As alternative solutions, the first is to express the traction transformer copper loss in the form of I²R, that is, to perform an equivalent substitution calculation by multiplying the square of the current at each moment by the equivalent resistance of the transformer; the second is to introduce weight coefficients for each item in the weighted summation of E_total to reflect the importance of different energy-saving types or differences in conversion.

[0058] Optionally, using the total energy saving calculation sub-model, the traction transformer loss optimization energy saving and total energy saving are calculated based on the photovoltaic energy saving, energy storage peak-valley arbitrage energy saving, energy storage network loss optimization energy saving, traction transformer loss coefficient, traction load value at each time, photovoltaic output value at each time, energy storage discharge power at each time, and energy storage charging power at each time. This includes: substituting the traction load value at each time into the loss calculation expression corresponding to the traction transformer loss coefficient to obtain the transformer loss value before connection at each time; subtracting the photovoltaic output value and energy storage discharge power at each time from the traction load value at each time and adding the energy storage charging power at each time, and then substituting this into the loss calculation expression to obtain the transformer loss value after connection at each time; using the difference between the transformer loss value before connection and the transformer loss value after connection at each time as the accumulated term, and accumulating it sequentially according to the time step within the evaluation period to obtain the traction transformer loss optimization energy saving; and summing the photovoltaic energy saving, energy storage peak-valley arbitrage energy saving, energy storage network loss optimization energy saving, and traction transformer loss optimization energy saving to obtain the total energy saving.

[0059] Substituting the traction load value P_L(t) at each time moment into the loss calculation expression corresponding to the traction transformer loss coefficient, we obtain the transformer loss value P_loss_before(t) before connection at each time moment. In one optional implementation, it can be calculated using the formula P_loss_before(t) = η_t·[P_L(t)]², where P_loss_before(t) is the traction transformer loss value before the photovoltaic-storage system is connected at time t, in kW; η_t is the traction transformer loss coefficient, which physically represents the ratio between the copper loss of the traction transformer and the square of the active power of the load it bears, and can be obtained by fitting the short-circuit loss parameters on the traction transformer nameplate or historical operating data. For example, η_t can be taken as η_t = Pk / Sn², where Pk is the rated short-circuit loss of the traction transformer (in kW), Sn is the rated capacity of the traction transformer (in kVA), and the specific value of η_t is not limited; P_L(t) is the traction load value at time t, in kW.

[0060] Subtracting the photovoltaic output and energy storage discharge power at each time from the traction load value at each time, and then adding the energy storage charging power at each time, the transformer loss value P_loss_after(t) after connection is obtained by substituting it into the same loss calculation expression. In an optional implementation, it can be calculated by the formula P_loss_after(t) = η_t·[P_L(t)−P_pv(t)−P_dis(t)+P_ch(t)]², where P_loss_after(t) is the traction transformer loss value after the photovoltaic-energy storage system is connected at time t, in kW; P_pv(t) is the photovoltaic output value at time t, in kW; P_dis(t) is the energy storage discharge power at time t, in kW; P_ch(t) is the energy storage charging power at time t, in kW; and η_t and P_L(t) have the same meaning as before. The physical meaning of the above formula is: after being connected to the photovoltaic-storage system, the actual equivalent load borne by the traction transformer is equal to the original traction load minus the photovoltaic output and energy storage discharge power, plus the energy storage charging power. Substituting this equivalent load into the loss expression, the transformer loss after connection can be obtained.

[0061] The difference between the transformer loss value before and after connection at each time point is used as the accumulator. This accumulator is then performed over a time step Δt within the evaluation period to obtain the optimized energy saving of the traction transformer, Wtr. In one optional implementation, it can be calculated using the formula Wtr = Σ_t[P_loss_before(t) − P_loss_after(t)]·Δt, where Wtr is the optimized energy saving of the traction transformer in kWh; Δt is the time step in hours (e.g., 1 hour, 15 minutes, 5 minutes); and the summation range of the summation symbol Σ_t covers the entire evaluation period, which can be a day, a month, a quarter, or a year, etc., without specific limitations.

[0062] The total energy saving E_total is obtained by summing the photovoltaic energy saving E_PV, the energy saving Warb from energy storage peak-valley arbitrage, the energy saving W_loss from energy storage network loss optimization, and the energy saving Wtr from traction transformer loss optimization. In one optional implementation, it can be calculated using the formula E_total = E_PV + Warb + W_loss + Wtr, where E_total is the total energy saving of the photovoltaic-storage traction power supply system, in kWh; E_PV is the photovoltaic energy saving, i.e., the energy saving converted from photovoltaic power generation itself, in kWh; Warb is the energy saving from energy storage peak-valley arbitrage, i.e., the equivalent energy saving converted from energy storage charging and discharging arbitrage using the peak-valley electricity price difference, in kWh; W_loss is the energy saving from energy storage network loss optimization, i.e., the reduction of line network losses by energy storage through optimizing feeder power flow distribution, in kWh; and Wtr has the same meaning as above.

[0063] S207. Determine the model grid power consumption based on multi-source basic data and traction transformer loss coefficient. Obtain the accuracy index based on the model grid power consumption and the actual grid power consumption. When the accuracy index is less than or equal to the preset accuracy threshold, obtain the evaluation result based on the total energy saving, photovoltaic energy saving, energy storage peak-valley arbitrage energy saving, energy storage grid loss optimization energy saving, and traction transformer loss optimization energy saving. Specifically, based on multi-source basic data and the traction transformer loss coefficient η_t, the power consumption W_model of the model grid is determined. In one optional implementation, the calculation can be performed using the formula W_model=Σ_t{[P_L(t)−P_pv(t)−P_dis(t)+P_ch(t)]+η_t·[P_L(t)−P_pv(t)−P_dis(t)+P_ch(t)]²}·Δt, where W_model is the model grid power consumption in kWh; P_L(t) is the traction load at time t in kW; P_pv(t) is the photovoltaic output at time t in kW; P_dis(t) is the energy storage discharge power at time t in kW; P_ch(t) is the energy storage charging power at time t in kW; Δt is the time step in hours; and the summation covers the entire evaluation period. The first term in the above formula is the active power corresponding to the equivalent load on the secondary side of the traction transformer, and the second term is the corresponding transformer copper power consumption. The sum of the two is the total grid power consumption drawn from the grid side. Then, the accuracy index ε is obtained based on the model grid power consumption W_model and the actual grid power consumption E_g. In one optional implementation, the relative error can be calculated as the accuracy index using the formula ε=|W_model−E_g| / E_g, where E_g is the actual grid power consumption obtained through electricity metering or dispatching system, in kWh, and ε is the dimensionless relative error value. As an alternative implementation, the root mean square error RMSE, mean absolute percentage error MAPE, or coefficient of determination R² can also be used as accuracy indices, without any specific limitation. Next, the accuracy index ε is compared with the preset accuracy threshold ε0 (e.g., 0.03, 0.05, or 0.10, the specific value of which can be configured according to the engineering evaluation requirements, and is not limited here): When ε≤ε0, the model accuracy is considered to meet the requirements. Then, based on the total energy saving E_total, photovoltaic energy saving E_PV, energy storage peak-valley arbitrage energy saving Warb, energy storage network loss optimization energy saving W_loss, and traction transformer loss optimization energy saving Wtr, the evaluation results are generated and output. The evaluation results may include the energy saving values ​​of each sub-item, the total energy saving value, the proportion of each sub-item energy saving in the total energy saving, the energy saving rate (i.e., the ratio of total energy saving to actual grid power consumption), and a visualization report of energy saving benefits presented in chart form, etc. When ε>ε0, the model accuracy is determined to not meet the requirements. The process can be repeated to re-identify and correct the traction transformer loss coefficient or other model parameters until the accuracy index meets the conditions. Thus, the accuracy and reliability of the evaluation results are guaranteed through the model self-verification mechanism.

[0064] Optionally, the model grid power consumption is determined based on multi-source basic data and the traction transformer loss coefficient. An accuracy index is obtained based on the model grid power consumption and the actual grid power consumption. When the accuracy index is less than or equal to a preset accuracy threshold, an evaluation result is obtained based on total energy savings, photovoltaic energy savings, energy storage peak-valley arbitrage energy savings, energy storage grid loss optimization energy savings, and traction transformer loss optimization energy savings. This includes: subtracting the photovoltaic output value and energy storage discharge power at each moment from the traction load value at each moment, and then adding the energy storage charging power at each moment, and comparing this with the traction transformer loss coefficient. The model grid power consumption is calculated at each moment and then accumulated over time within the evaluation period to obtain the model grid power consumption. The absolute value of the difference between the model grid power consumption and the actual grid power consumption is divided by the actual grid power consumption to obtain the accuracy index. The accuracy index is compared with a preset accuracy threshold. When the accuracy index is less than or equal to the preset accuracy threshold, the total energy saving, photovoltaic energy saving, energy storage peak-valley arbitrage energy saving, energy storage grid loss optimization energy saving, and traction transformer loss optimization energy saving are classified and combined according to preset evaluation dimensions to obtain the evaluation result.

[0065] Specifically, the traction load value P_L(t) at each time moment is subtracted from the photovoltaic output value P_pv(t) and the energy storage discharge power P_dis(t) at each time moment, and the energy storage charging power P_ch(t) at each time moment is added together. Then, this is used to calculate the model grid power value P_model(t) at each time moment with the traction transformer loss coefficient η_t. In one optional implementation, it can be calculated using the formula P_model(t) = [P_L(t) − P_pv(t) − P_dis(t) + P_ch(t)] + η_t·[P_L(t) − P_pv(t) − P_dis(t) + P_ch(t)]², where P_model(t) is the model grid power value at time t, in kW; P_L(t) is the traction load value at time t, in kW; P_pv(t) is the photovoltaic output value at time t, in kW; P_d... is(t) is the energy storage discharge power at time t, in kW; P_ch(t) is the energy storage charging power at time t, in kW; η_t is the traction transformer loss coefficient, which is the proportionality coefficient between the copper loss of the traction transformer and the square of the active power of the load it bears, in 1 / kW; the first term in the above formula represents the power drawn from the grid side corresponding to the equivalent active load on the secondary side of the traction transformer, and the second term represents the copper loss power generated by the equivalent load on the traction transformer. The sum of the two is the total active power drawn from the grid side at that time.

[0066] The grid power consumption value P_model(t) at each time point within the evaluation period is summed over time to obtain the grid power consumption W_model. In one optional implementation, it can be calculated using the formula W_model = Σ_t·P_model(t)·Δt, where W_model is the grid power consumption in kWh; Δt is the time step in hours (h), such as 1 hour, 15 minutes, or 5 minutes, etc.; and the summation range of the summation symbol Σ_t covers the entire evaluation period, which can be a day, a month, a quarter, or a year, etc., without specific limitations.

[0067] The accuracy index ε is obtained by dividing the absolute value of the difference between the model's grid power consumption W_model and the actual grid power consumption E_g by the actual grid power consumption E_g. In one optional implementation, it can be calculated using the formula ε = |W_model−E_g| / E_g, where ε is a dimensionless accuracy index used to characterize the degree of deviation between the model's output grid power consumption and the actual value; E_g is the actual grid power consumption obtained through electricity metering or a dispatching system, in kWh; and |·| represents the absolute value operation. The smaller ε is, the closer the model output is to the actual value, and the higher the model accuracy.

[0068] The accuracy index ε is compared with a preset accuracy threshold ε0, where ε0 is a pre-set upper limit of the allowable relative error, such as 0.03, 0.05, or 0.10. Its specific value can be configured according to the accuracy requirements of the engineering evaluation and is not limited here. When ε≤ε0, the model accuracy is considered to meet the requirements, and the evaluation result generation stage begins. When ε>ε0, the model accuracy is determined to be unsatisfactory, and the process can return to re-identify and correct the traction transformer loss coefficient η_t or other model parameters before re-executing the aforementioned steps, until the accuracy index meets the conditions.

[0069] When the accuracy index ε is less than or equal to the preset accuracy threshold ε0, the total energy saving E_total, photovoltaic energy saving E_PV, energy storage peak-valley arbitrage energy saving Warb, energy storage network loss optimization energy saving W_loss, and traction transformer loss optimization energy saving Wtr are combined according to the preset evaluation dimensions to obtain the evaluation result. In one optional implementation, the preset evaluation dimensions may include, but are not limited to: classifying by energy-saving source dimension, classifying E_PV as photovoltaic power generation energy saving, Warb as energy storage operation strategy energy saving, and W_loss and Wtr as electrical loss optimization energy saving; classifying by time scale dimension, summarizing and statistically analyzing the energy saving of each sub-item by day, month, quarter, and year; classifying by energy saving percentage dimension, calculating the percentage of each sub-item's energy saving in the total energy saving E_total, i.e., η_i=W_i / E_total×100%, where W_i is the energy saving of the i-th sub-item, and η_i is its percentage in the total energy saving; and classifying by energy saving rate dimension, calculating the energy saving rate η=E_total / E_g×100%, which is used to characterize the overall energy saving level.

[0070] S208. Based on the proportion of energy saving of each sub-item to the total energy saving in the evaluation results, the energy storage charging and discharging power correction value and the photovoltaic inverter reactive power output setting value are matched from the preset strategy library and sent to the energy storage converter and photovoltaic inverter to control the photovoltaic-storage microgrid to operate according to the corrected parameters.

[0071] Based on the evaluation results output in step S207, the proportions of photovoltaic energy saving E_PV, energy storage peak-valley arbitrage energy saving Warb, energy storage network loss optimization energy saving W_loss, and traction transformer loss optimization energy saving Wtr in the total energy saving E_total are calculated according to the formula η_i=W_i / E_total×100%, where η_i is the percentage of the i-th sub-item's energy saving in the total energy saving, and W_i is the corresponding sub-item's energy saving value in kWh. Then, the feature vector [η_PV, η_arb, η_loss, η_tr] formed by the above proportions is used as the query key and input into the preset strategy library for matching to obtain the energy storage charging and discharging power correction value ΔP_bess and the photovoltaic inverter reactive power output setting value Q_pv. The preset strategy library can be stored in advance in the form of key-value pairs in a relational database, in-memory database, or local configuration. Each record in the file contains a set of percentage interval conditions and their corresponding ΔP_bess and Q_pv. For example, when η_arb is low and η_loss is high, the matching is to increase ΔP_bess during peak periods and increase charging during valley periods (the amplitude can be ±5% to ±15% of the rated power) while keeping Q_pv unchanged. When η_tr is high, the matching is to reduce the amplitude of charging and discharging power and moderately increase Q_pv within the inverter capacity margin to support the bus voltage. Specific rules can be pre-configured based on engineering experience and are not limited here. The matching method can use rule-based conditional branching or decision trees for precise matching, or it can use nearest neighbor retrieval based on Euclidean distance or cosine similarity, or use a pre-trained classification model (such as random forest, XGBoost, etc.) to directly output the strategy number with the feature vector as input. The specific method is not limited. After matching, ΔP_bess and Q_pv are encapsulated in accordance with IEC Control messages using communication protocols such as 61850 MMS, Modbus TCP, or DNP3.0 are sent to the energy storage converter and photovoltaic inverter via the southbound communication interface of the photovoltaic-storage microgrid energy management system. The energy storage converter updates its internal power reference value according to ΔP_bess, i.e., P_ref_new = P_ref_old + ΔP_bess, and performs charging and discharging actions according to the new reference values ​​under the constraints of the upper and lower limits of the energy storage state of charge SOCmax and SOCmin, and the rated charging and discharging power Pchn and Pdisn. The photovoltaic inverter adjusts the setpoint of its reactive power control loop according to Q_pv, and injects or absorbs reactive power into the traction power supply bus according to the new reactive power setpoint while ensuring the priority output of active power. The above parameter correction cycle can be configured to be executed once a day, once a week, or once a month according to actual engineering needs, or it can be triggered for immediate correction when the accuracy index ε exceeds the threshold for multiple consecutive cycles.

[0072] This embodiment also discloses an electronic device, as shown in the reference. Figure 3The electronic device may include: at least one processor 301, at least one communication bus 302, a user interface 303, a network interface 304, and at least one memory 305. The communication bus 302 is used to enable communication between these components. The user interface 303 may include a display screen or a camera; optionally, the user interface 303 may also include a standard wired interface or a wireless interface. The network interface 304 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0073] The processor 301 may include one or more processing cores. The processor 301 connects to various parts of the server using various interfaces and lines, and performs various server functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in memory 305, and by calling data stored in memory 305. Optionally, the processor 301 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 301 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required for display; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the processor 301 and may be implemented as a separate chip.

[0074] The memory 305 may include random access memory (RAM) or read-only memory. Optionally, the memory 305 may include a non-transitory computer-readable storage medium. The memory 305 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 305 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 305 may also be at least one storage device located remotely from the aforementioned processor 301. Figure 3 As shown, the memory 305, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an application program for an energy assessment method for connecting a photovoltaic-storage microgrid to an electrified railway.

[0075] exist Figure 3 In the electronic device shown, the user interface 303 is mainly used to provide an input interface for the user and to obtain the user input data; while the processor 301 can be used to call the application program stored in the memory 305, which is a method for assessing energy savings after a photovoltaic-storage microgrid is connected to an electrified railway. When executed by one or more processors 301, the electronic device performs one or more methods as described in the above embodiments.

[0076] The foregoing description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Other embodiments of this disclosure will be readily apparent to those skilled in the art upon consideration of the disclosure in this specification. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described in this disclosure. The specification and embodiments are to be considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.

Claims

1. A method for assessing energy savings after a photovoltaic-storage microgrid is connected to an electrified railway, characterized in that, The method includes: The evaluation boundary is determined, and basic parameters are set based on the evaluation boundary. The basic parameters include time step, evaluation period, traction transformer loss coefficient, energy storage charging and discharging efficiency, and time period division results of railway grid electricity price. According to the time step, multi-source basic data is collected within the evaluation period. The multi-source basic data includes traction load value, photovoltaic output value, energy storage charging power, energy storage discharging power and grid power metering value at each time. The grid power metering values ​​at each time within the evaluation period are accumulated to obtain the actual grid power consumption. Based on the multi-source basic data, the loss coefficient of the traction transformer, the energy storage charging and discharging efficiency, and the time period division results, a photovoltaic-storage synergistic energy-saving assessment model is constructed. The photovoltaic-storage synergistic energy-saving assessment model includes a photovoltaic energy-saving calculation sub-model, an energy storage energy-saving calculation sub-model, and a total energy-saving calculation sub-model. The photovoltaic energy saving calculation sub-model calculates the photovoltaic energy saving based on the photovoltaic output value, traction load value, and energy storage discharge power at each time. The energy-saving calculation sub-model for energy storage calculates the energy-saving energy from peak-valley arbitrage and the energy-saving energy from energy storage network loss optimization based on the multi-source basic data, the energy storage charging and discharging efficiency, and the time period division results. Using the total energy saving calculation sub-model, the energy saving of photovoltaic power, the energy saving of energy storage peak-valley arbitrage, the energy saving of energy storage network loss optimization, the loss coefficient of traction transformer, the traction load value at each time, the photovoltaic output value at each time, the energy storage discharge power at each time, and the energy storage charging power at each time are used to calculate the energy saving of traction transformer loss optimization and the total energy saving. The model grid power consumption is determined based on the multi-source basic data and the traction transformer loss coefficient. An accuracy index is obtained based on the model grid power consumption and the actual grid power consumption. When the accuracy index is less than or equal to a preset accuracy threshold, an evaluation result is obtained based on the total energy saving, the photovoltaic energy saving, the energy saving from energy storage peak-valley arbitrage, the energy saving from energy storage grid loss optimization, and the energy saving from traction transformer loss optimization. Based on the proportion of energy saving in each sub-item to the total energy saving in the assessment results, the energy storage charging and discharging power correction value and the photovoltaic inverter reactive power output setting value are matched from the preset strategy library and sent to the energy storage converter and photovoltaic inverter to control the photovoltaic-storage microgrid to operate according to the corrected parameters.

2. The method according to claim 1, characterized in that, The photovoltaic-storage synergistic energy-saving assessment model is constructed based on the multi-source basic data, the traction transformer loss coefficient, the energy storage charging and discharging efficiency, and the time period division results, including: For the missing moments in the multi-source basic data, interpolation is performed using the arithmetic mean of adjacent moments to fill in the missing moments. After removing samples that exceed a preset fluctuation threshold according to a preset dispersion criterion, the missing moments are filled in again to obtain the corrected multi-source basic data. Based on the time-series matching relationship between the photovoltaic output value, the energy storage discharge power and the traction load value at each time in the corrected multi-source basic data, dynamic calculation rules are constructed, and the dynamic calculation rules are encapsulated into the photovoltaic energy saving calculation sub-model. The photovoltaic local consumption rate at each moment output by the photovoltaic energy saving calculation sub-model is used as the criterion. Combined with the time period division results and the conversion relationship between the energy storage charging and discharging efficiency and the charging and discharging energy, a joint operation rule is constructed, and the joint operation rule is encapsulated into the energy storage energy saving calculation sub-model. The traction transformer loss coefficient is decoupled into no-load loss component and load loss component. Based on the no-load loss component and the load loss component, and combined with the equivalent load smoothing result output by the energy storage energy saving calculation sub-model, a summation operation rule is constructed, and the summation operation rule is encapsulated into the total energy saving calculation sub-model. The photovoltaic-storage synergistic energy-saving assessment model is obtained by cascading and coupling the input-output dependencies between the photovoltaic energy-saving calculation sub-model, the energy storage energy-saving calculation sub-model, and the total energy-saving calculation sub-model.

3. The method according to claim 2, characterized in that, The photovoltaic energy saving calculation sub-model calculates the photovoltaic energy saving based on the photovoltaic output value, traction load value, and energy storage discharge power at each time point, including: The photovoltaic output value at each time moment is added to the energy storage discharge power at each time moment to obtain the comprehensive photovoltaic power supply at each time moment. The photovoltaic power supply at each time moment is compared with the traction load value at each time moment. When the photovoltaic power supply at each time moment is less than or equal to the traction load value at each time moment, the photovoltaic local consumption rate at each time moment is assigned a preset upper limit value of the consumption rate. When the total power supplied by the photovoltaic side at each time is greater than the traction load value at each time, the ratio of the traction load value at each time to the total power supplied by the photovoltaic side at each time is taken as the local photovoltaic consumption rate at each time. The photovoltaic energy saving at each moment is obtained by multiplying the photovoltaic output value at each moment, the photovoltaic local consumption rate at each moment, and the time step. The photovoltaic energy saving is obtained by summing the photovoltaic energy saving at each moment within the evaluation period.

4. The method according to claim 1, characterized in that, The energy-saving calculation sub-model for energy storage calculates the energy-saving energy from peak-valley arbitrage and the energy-saving energy from network loss optimization based on the multi-source basic data, the energy storage charging and discharging efficiency, and the time period division results. This includes: According to the time period division results, each moment in the evaluation period is divided into valley moment, flat moment and peak moment. The energy storage charging power of each moment is determined at the valley moment, the energy storage discharging power of each moment is determined at the peak moment, and the energy storage charging power and energy storage discharging power of each moment are set to zero at the flat moment. Based on the calculation relationship between the energy storage discharge power, the energy storage charging power, the energy storage charging and discharging efficiency, and the time step, the energy storage peak-valley arbitrage energy saving at each moment is obtained, and the energy storage peak-valley arbitrage energy saving at each moment within the evaluation period is accumulated in time to obtain the energy storage peak-valley arbitrage energy saving. Based on the ratio of the traction load value at each moment to the preset power supply voltage level, the equivalent current before connection at each moment is obtained; Based on the calculation relationship between the traction load value, photovoltaic output value, energy storage discharge power, energy storage charging power and the preset power supply voltage level at each time, the equivalent current after connection at each time is obtained. Based on the calculation relationship between the square of the equivalent current before access at each time point, the square of the equivalent current after access at each time point, the preset equivalent resistance of the contact network, and the time step, the energy storage network loss optimization energy saving is obtained by performing time-series accumulation within the evaluation period.

5. The method according to claim 4, characterized in that, Determining the energy storage charging power at each time point during the valley period and determining the energy storage discharging power at each time point during the peak period includes: At the valley time, the energy storage charging power at each time is obtained based on the preset rated energy storage charging power and combined with the preset upper limit of energy storage state of charge. At the peak time, the energy storage discharge power at each time is obtained by taking the preset rated discharge power of energy storage as a reference and combining the preset limit of energy storage state of charge. The energy storage state of charge value at each moment is obtained by performing time-series iterative calculations using the energy storage charging power, energy storage discharging power, energy storage charging and discharging efficiency, and time step.

6. The method according to claim 1, characterized in that, The total energy saving calculation sub-model calculates the traction transformer loss optimization energy saving and the total energy saving based on the photovoltaic energy saving, the energy saving from energy storage peak-valley arbitrage, the energy saving from energy storage network loss optimization, the traction transformer loss coefficient, the traction load value at each time moment, the photovoltaic output value at each time moment, the energy storage discharge power at each time moment, and the energy storage charging power at each time moment. This includes: Substitute the traction load values ​​at each time point into the loss calculation expression corresponding to the traction transformer loss coefficient to obtain the transformer loss values ​​before connection at each time point. Subtract the photovoltaic output value and the energy storage discharge power at each time from the traction load value at each time, and add the energy storage charging power at each time, then substitute the result into the loss calculation expression to obtain the transformer loss value after connection at each time. The difference between the transformer loss value before access at each time point and the transformer loss value after access at each time point is used as the accumulator. The accumulator is performed sequentially according to the time step within the evaluation period to obtain the energy saving of the traction transformer loss optimization. The total energy saving is obtained by summing the energy saving from photovoltaic power generation, the energy saving from energy storage peak-valley arbitrage, the energy saving from optimized energy storage network loss, and the energy saving from optimized traction transformer loss.

7. The method according to claim 6, characterized in that, The model grid power consumption is determined based on the multi-source basic data and the traction transformer loss coefficient. An accuracy index is obtained based on the model grid power consumption and the actual grid power consumption. When the accuracy index is less than or equal to a preset accuracy threshold, an evaluation result is obtained based on the total energy saving, the photovoltaic energy saving, the energy saving from energy storage peak-valley arbitrage, the energy saving from energy storage grid loss optimization, and the energy saving from traction transformer loss optimization. This evaluation includes: The photovoltaic output value and the energy storage discharge power at each time are deducted from the traction load value at each time, and the energy storage charging power at each time is added together. Then, the model grid power value at each time is calculated with the loss coefficient of the traction transformer. The model grid power value at each time is then accumulated over time within the evaluation period to obtain the model grid power consumption. The accuracy index is obtained by dividing the absolute value of the difference between the model grid power consumption and the actual grid power consumption by the actual grid power consumption. The accuracy index is compared with the preset accuracy threshold. When the accuracy index is less than or equal to the preset accuracy threshold, the total energy saving, the photovoltaic energy saving, the energy saving from energy storage peak-valley arbitrage, the energy saving from energy storage network loss optimization, and the energy saving from traction transformer loss optimization are combined according to preset evaluation dimensions to obtain the evaluation result.

8. An electronic device, characterized in that, The device includes a processor, a memory, a user interface, and a network interface. The memory is used to store instructions. The user interface and the network interface are both used to communicate with other devices. The processor is used to execute the instructions stored in the memory to cause the electronic device to perform the method as described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed, perform the method as described in any one of claims 1-7.

10. A computer program product, characterized in that, When the computer program product is run on an electronic device, it causes the electronic device to perform the method as described in any one of claims 1-7.

Images