Retired battery photovoltaic on-site consumption closed-loop scheduling method and system

Abstract

The application discloses a kind of retired battery photovoltaic in-place consumption closed-loop scheduling method and system, belong to new energy storage technology field, including: the operation data of retired battery, photovoltaic, local load, grid-connected point and power conversion equipment are collected;According to the health bearing capacity index of retired battery health degradation characteristics and generate safe charge-discharge power boundary is calculated;According to the real-time power balance relationship of power supply-storage-use and power grid is established by trusted photovoltaic available power and local load forecast power, under the constraints of battery health boundary, state of charge, temperature, equipment power and grid safety, the rolling scheduling model is solved, and the battery charge-discharge power instruction and photovoltaic limit power instruction are obtained;The corresponding instruction is issued to energy storage converter and photovoltaic inverter, and the next scheduling cycle is recalculated according to actual execution power, battery state and grid-connected point feedback.The application can improve the in-place consumption capacity of photovoltaic under the premise of guaranteeing the safe operation of retired battery and grid-connected.

Description

Technical Field

[0001] This invention belongs to the fields of new energy storage, cascade utilization of retired batteries and energy management and control technology, specifically involving a closed-loop scheduling method and system for local photovoltaic consumption of retired batteries. Background Technology

[0002] With the development of the new energy vehicle industry, the number of retired power batteries continues to increase. Although some retired power batteries can no longer meet the high-rate operating requirements of vehicles, they still have remaining capacity that can be used for distributed photovoltaic energy storage, user-side energy storage, and microgrid energy storage. In the scenario of local photovoltaic consumption, photovoltaic output and local load are constantly changing. The energy management system needs to periodically determine the charging and discharging power of retired batteries to keep the photovoltaic power generation, retired battery charging and discharging power, local load power, and grid connection point exchange power in balance.

[0003] However, retired power batteries vary significantly in their origin, service history, and degree of degradation. Existing energy management systems typically schedule batteries based on rated power, fixed power limits, or a single health status indicator. Capacity retention or health status primarily reflects the decay of usable capacity and cannot simultaneously characterize the impact of internal resistance growth, energy efficiency decline, accelerated temperature rise, and abnormal voltage response on instantaneous and sustained power carrying capacity. Therefore, even if two retired batteries have similar capacity retention rates, their actual safe charge and discharge power may differ significantly. Using fixed power boundaries can easily lead to scheduling commands exceeding the actual carrying capacity of high-risk batteries, or prevent batteries in better condition from fully participating in photovoltaic power consumption.

[0004] On the other hand, in existing solutions, battery health assessment and energy storage scheduling are usually carried out separately: the health assessment module outputs the health level, while the scheduling module still calculates the charging and discharging power based on fixed power limits or empirical thresholds. The health assessment results do not directly form a scheduling feasible domain. Although some solutions calculate the curtailed power or purchased power, they do not clearly define the power balance relationship between photovoltaics, loads, energy storage and grid connection points, nor do they jointly constrain the allowed purchased power, allowed reverse power and power change rate of the grid connection point.

[0005] Furthermore, calculating only the theoretically optimal charge and discharge power without considering the actual execution results of the energy storage converter cannot handle operational deviations such as power point tracking errors, sudden temperature rises, grid connection point power exceeding limits, or battery management system alarms. Therefore, a closed-loop scheduling method and system is developed that combines the health carrying capacity of retired batteries, photovoltaic and load forecasting, power balance, grid connection safety constraints, power command execution, and state feedback. This allows the battery health status to directly limit the executable power, and the execution results can be used for recalculation in the next scheduling cycle. Summary of the Invention

[0006] This invention aims to address the following problems in existing retired battery energy storage scheduling: the disconnect between health evaluation results and charge / discharge power commands; the lack of explicit power balance constraints between photovoltaic power generation and local load power; the absence of grid connection point safety boundaries in joint optimization; and the lack of feedback correction between theoretical power commands and actual execution status.

[0007] To achieve the above objectives, this invention provides a closed-loop scheduling method for the local consumption of retired photovoltaic cells, which executes the following steps according to a continuous scheduling cycle: S1: Collect operational data from the retired battery side, photovoltaic power generation side, local load side, grid connection point, and power conversion equipment side; among which, the operational data from the retired battery side includes battery terminal voltage, operating current, state of charge, battery body temperature, and battery management system alarm information; the operational data from the photovoltaic power generation side includes actual photovoltaic output power, and at least one of solar irradiance, photovoltaic module temperature, and weather forecast data; the operational data from the local load side includes local load power; and the operational data from the grid connection point and power conversion equipment side includes grid connection point exchange power, maximum allowable power purchase capacity, maximum allowable reverse power supply capacity, maximum grid connection point power change rate, and maximum output power of the energy storage converter. S2: Extract multidimensional health degradation features from the operating data of retired batteries, calculate the health carrying capacity index of retired batteries, and generate the safe charging power boundary and safe discharging power boundary for the current scheduling cycle based on the health carrying capacity index of retired batteries. S3: Based on the photovoltaic physical model, historical photovoltaic prediction errors, and the degree of irradiance variation, obtain the reliable photovoltaic available power in the scheduling prediction time domain, and obtain the local load prediction power in the scheduling prediction time domain; S4: Using the charging and discharging power of retired batteries, the curtailed power of photovoltaics, and the exchange power at the grid connection point as scheduling variables, establish the real-time power balance relationship between the generation-storage-consumption and the grid among the reliable available photovoltaic power, the curtailed power, the exchange power at the grid connection point, the local load power, and the charging and discharging power of retired batteries. Based on the safe charging power boundary, safe discharging power boundary, state of charge constraint, temperature constraint, power constraint of energy storage converter, and grid security constraint, a scheduling feasible domain is constructed. The grid security constraint includes the exchange power constraint at the grid connection point and the power change rate constraint at the grid connection point. S5: Within the feasible scheduling domain, with the optimization objectives of photovoltaic curtailment power generation, grid power purchase, degradation cost of retired batteries, charging and discharging power switching cost, photovoltaic prediction uncertainty risk and grid connection point power fluctuation, the retired battery charging and discharging power sequence and photovoltaic curtailment power sequence in the rolling optimization time domain are obtained. The rolling optimization time domain consists of multiple prediction steps, and the power value corresponding to the first prediction step in the two sequences is selected as the charging and discharging power command and photovoltaic curtailment power command for the current scheduling cycle, respectively. S6: The energy storage converter receives the charging and discharging power command for the current scheduling cycle, and the photovoltaic inverter receives the actual charging and discharging power of the retired battery. When the deviation between the actual charging and discharging power of the retired battery and the charging and discharging power command for the current scheduling cycle exceeds a preset deviation threshold, or the grid connection point exchange power exceeds the allowable range, or the battery management system generates an alarm, the abnormal operation level of the retired battery is determined based on the deviation between the actual charging and discharging power of the retired battery and the charging and discharging power command for the current scheduling cycle, or the grid connection point exchange power, or the alarm information generated by the battery management system. The abnormality level includes minor abnormality, moderate abnormality, or severe abnormality. When the retired battery is operating under a minor abnormality, the charging and discharging power command for the current scheduling cycle is limited and corrected. When the retired battery is operating under a moderate abnormality, S2 to S5 are executed based on the latest measured state, updating the health carrying capacity index, safe charging and discharging power boundary, and scheduling power command of the retired battery, and entering the next scheduling cycle. When the retired battery is operating under a severe abnormality, the retired battery exits the current charging and discharging scheduling cycle and the charging and discharging power command for the current scheduling cycle is set to zero.

[0008] According to another aspect of the present invention, a closed-loop scheduling system for on-site consumption of retired photovoltaic cells is provided, comprising a data sensing module, a health boundary generation module, a reliable prediction and power balance module, a rolling scheduling optimization module, a control execution module, and a feedback correction module.

[0009] According to another aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, wherein when the computer program is executed by a processor, the above-described closed-loop scheduling method for on-site consumption of retired photovoltaic cells based on dynamic boundaries of health carrying capacity is implemented.

[0010] In summary, the above-described technical solutions conceived in this invention can achieve the following beneficial effects: (1) The present invention directly converts the health carrying capacity index formed by capacity, internal resistance, energy efficiency, temperature rise rate and voltage response deviation into a safe charging and discharging power boundary, and uses the boundary as a hard constraint for rolling scheduling, so that the health evaluation results will inevitably participate in the power command generation of each scheduling cycle.

[0011] (2) This invention establishes an explicit power balance relationship between reliable photovoltaic available power, photovoltaic power limit, grid connection point exchange power, local load power and retired battery charging and discharging power, so that the scheduling results correspond to the actual executable power allocation of power generation, energy storage, power consumption and grid.

[0012] (3) The present invention uses the allowable power purchase, allowable reverse power and power change rate at the grid connection point as constraints of the dispatchable feasible domain, so as to improve the local photovoltaic consumption rate and avoid the power limit exceeding or sudden change at the grid connection point.

[0013] (4) The present invention conservatively corrects the available photovoltaic power based on the reliability of photovoltaic prediction and sets a prediction uncertainty risk term in the objective function, thereby reducing the impact of photovoltaic prediction error on the power command of retired batteries and the power of grid connection point.

[0014] (5) The present invention sends the charge and discharge power command and the photovoltaic power limit command after safety limit to the energy storage converter and the photovoltaic inverter respectively, and corrects or re-solves according to the actual power, battery status, grid connection point exchange power and battery management system alarm, forming a closed-loop control across scheduling cycles. Attached Figure Description

[0015] Figure 1 The flowchart illustrates the closed-loop scheduling method for on-site consumption of decommissioned photovoltaic cells based on dynamic boundaries of health carrying capacity, as provided in this embodiment of the invention.

[0016] Figure 2 This is a schematic diagram of the closed-loop scheduling system for on-site consumption of decommissioned photovoltaic cells based on dynamic boundaries of health carrying capacity, provided in an embodiment of the present invention.

[0017] Figure 3 This is a schematic diagram of the closed-loop scheduling device and computer-readable storage medium provided in an embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0019] In this invention, the terms "first," "second," etc. (if present) in the invention and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0020] Example 1: This invention discloses a closed-loop scheduling method for on-site consumption of decommissioned photovoltaic cells based on dynamic boundaries of health carrying capacity. For example... Figure 1 As shown, this method is based on the dynamic boundary of health carrying capacity and is executed cyclically according to a continuous scheduling cycle. It includes multi-source operation data acquisition, generation of health carrying capacity and safe power boundary, photovoltaic and load forecasting, power balance and scheduling feasible region construction, rolling optimization solution, power command issuance and execution feedback, and specifically includes the following steps: S1: Collect operational data from the retired battery side, photovoltaic power generation side, local load side, grid connection point, and power conversion equipment side; among which, the operational data from the retired battery side includes battery terminal voltage, operating current, state of charge, battery body temperature, and battery management system alarm information; the operational data from the photovoltaic power generation side includes actual photovoltaic output power, and at least one of solar irradiance, photovoltaic module temperature, and weather forecast data; the operational data from the local load side includes local load power; and the operational data from the grid connection point and power conversion equipment side includes grid connection point exchange power, maximum allowable power purchase capacity, maximum allowable reverse power supply capacity, maximum grid connection point power change rate, and maximum output power of the energy storage converter. S2: Extract multidimensional health degradation features from the operating data of retired batteries, calculate the health carrying capacity index of retired batteries, and generate the safe charging power boundary and safe discharging power boundary for the current scheduling cycle based on the health carrying capacity index of retired batteries. S3: Based on the photovoltaic physical model, historical photovoltaic prediction errors, and the degree of irradiance variation, obtain the reliable photovoltaic available power in the scheduling prediction time domain, and obtain the local load prediction power in the scheduling prediction time domain; S4: Using the charging and discharging power of retired batteries, the curtailed power of photovoltaics, and the exchange power at the grid connection point as scheduling variables, establish the real-time power balance relationship between the generation-storage-consumption and the grid among the reliable available photovoltaic power, the curtailed power, the exchange power at the grid connection point, the local load power, and the charging and discharging power of retired batteries. Based on the safe charging power boundary, safe discharging power boundary, state of charge constraint, temperature constraint, power constraint of energy storage converter, and grid security constraint, a scheduling feasible domain is constructed. The grid security constraint includes the exchange power constraint at the grid connection point and the power change rate constraint at the grid connection point. S5: Within the feasible scheduling domain, with the optimization objectives of photovoltaic curtailment power generation, grid power purchase, degradation cost of retired batteries, charging and discharging power switching cost, photovoltaic prediction uncertainty risk, and grid connection point power fluctuation, the retired battery charging and discharging power sequence and the photovoltaic curtailment power sequence within the rolling optimization time domain are obtained. The rolling optimization time domain consists of multiple prediction step sizes, and the power value corresponding to the first prediction step size in the two sequences is selected as the charging and discharging power command and the photovoltaic curtailment power command for the current scheduling cycle, respectively. It should be noted that the feasible scheduling domain here refers to the power value range jointly defined by the retired battery safe charging power boundary, safe discharging power boundary, state of charge constraint, temperature constraint, energy storage converter power constraint, and grid security constraint.

[0021] S6: The energy storage converter receives the charging and discharging power command for the current scheduling cycle, and the photovoltaic inverter receives the actual charging and discharging power of the retired battery. When the deviation between the actual charging and discharging power of the retired battery and the charging and discharging power command for the current scheduling cycle exceeds a preset deviation threshold, or the grid connection point exchange power exceeds the allowable range, or the battery management system generates an alarm, the abnormal operation level of the retired battery is determined based on the deviation between the actual charging and discharging power of the retired battery and the charging and discharging power command for the current scheduling cycle, or the grid connection point exchange power, or the alarm information generated by the battery management system. The abnormality level includes minor abnormality, moderate abnormality, or severe abnormality. When the retired battery is operating under a minor abnormality, the charging and discharging power command for the current scheduling cycle is limited and corrected. When the retired battery is operating under a moderate abnormality, S2 to S5 are executed based on the latest measured state, updating the health carrying capacity index, safe charging and discharging power boundary, and scheduling power command of the retired battery, and entering the next scheduling cycle. When the retired battery is operating under a severe abnormality, the retired battery exits the current charging and discharging scheduling cycle and the charging and discharging power command for the current scheduling cycle is set to zero. It should be noted that the preset deviation threshold here refers to the first-level deviation threshold and the second-level deviation threshold used to distinguish different degrees of power tracking deviation; the allowable range in the grid connection point exchange power exceeding the allowable range refers to the power range limited by the upper limit of the allowable power purchase power and the upper limit of the allowable reverse power of the grid connection point; the limit correction refers to the limit of the charging and discharging power command of the current dispatch cycle with the corrected safe charging and discharging power range, and the photovoltaic power limiting command is recalculated in conjunction.

[0022] Specifically, at the beginning of each scheduling cycle, the current safe power boundary is first calculated using the latest collected data on retired batteries, photovoltaics, local loads, grid connection points, and power conversion equipment; then, the safe power boundary and forecast data are used to solve for the future... The scheduling power sequence is predicted in steps; only the first power value in the power sequence is issued for execution; at the end of the scheduling cycle, the actual running status is collected, and the model input for the next scheduling cycle is updated with the actual running status, thus forming a rolling closed loop.

[0023] Among them, the data collected from the retired battery side includes battery terminal voltage. Operating current State of charge Battery body temperature Cumulative number of loops Battery Management System (BMS) alarm information and actual charging and discharging power of the energy storage converter. Optionally, the cumulative number of loops can also be collected. This is used to correct health degradation characteristics. In this invention, the charging direction of retired batteries is defined as the positive current direction, i.e. >0 indicates charging. <0 indicates discharge; the sign convention for the charge and discharge power of retired batteries is the same.

[0024] Data collected from the photovoltaic power generation side includes solar irradiance. Photovoltaic module temperature Actual output power of photovoltaic inverter Inverter operating efficiency and future weather forecast data; local load-side data includes local real-time load power. Based on historical load data and the latest load data, the local load forecast power in the scheduling forecast time domain is obtained.

[0025] Data collected at the grid connection point and power conversion equipment side includes the actual switching power at the grid connection point. Maximum allowable power purchase capacity at grid connection point Maximum allowable reverse power Upper limit of power change rate at grid connection point Reference value of grid connection point switching power Maximum output power of energy storage converter Battery safety voltage range , Maximum allowable current Safe operating range under charged state [ , and the battery's maximum safe temperature .

[0026] Furthermore, the continuous operation process is divided into multiple scheduling cycles, the first... scheduling cycles Defined as: , in, For the first One scheduling cycle, For the first The start time of each scheduling cycle, The length of the scheduling period. This is the scheduling cycle number. In the [number]th [cycle / period]... At the start of each scheduling cycle, with The actual state at any given time is the initial state, and it is set to include... Rolling prediction time domain with prediction steps: , in, For the first The rolling prediction time domain corresponding to each scheduling cycle The rolling prediction time domain is the number of prediction steps included in the rolling prediction time domain; the rolling prediction time domain starts from the first prediction step after the current scheduling period and continues to the next prediction step. The prediction step ends.

[0027] Solve within the k-th scheduling period The power sequence is scheduled within the range, but only the power instruction corresponding to the first prediction step is executed; upon entering the next step... After one scheduling cycle, the system will be re-established based on the latest actual state. And solve it again.

[0028] The data from each side are processed by time alignment, outlier identification and removal, filtering, and unit unification to obtain the synchronous operation dataset for the k-th scheduling cycle: , in, For the first Synchronous running dataset for each scheduling cycle; and These are the terminal voltage and operating current of the retired battery, respectively. State of charge of retired batteries; This refers to the battery body temperature. and These are solar irradiance and photovoltaic module temperature, respectively. and These are the actual output power of the photovoltaic system and the local load power, respectively. and These are the actual switching power at the grid connection point and the actual charging and discharging power of the energy storage converter, respectively. This indicates an alarm status in the battery management system. Rated power, safe voltage range, maximum allowable current, safe operating range under state of charge, maximum safe temperature, and grid connection point power limit are stored separately as constraint parameters and are not included in standardization.

[0029] During the secondary utilization of retired batteries, in addition to capacity decay, there are also increased internal resistance, decreased charge and discharge efficiency, accelerated temperature rise, and abnormal voltage response. To obtain a health carrying capacity index characterizing the current power carrying capacity of retired batteries, based on the k-th scheduling cycle... Extract the following multidimensional health degradation features: (1) : , in, Indicates the first Estimated battery capacity for each scheduling cycle. This indicates the corresponding scheduling period, and I(t) represents the current acquisition value. This represents the change in state of charge during the scheduling period. To prevent the default minimum positive number with a denominator of zero, the integral sign is indicated in... The absolute value of the current is integrated over time.

[0030] Since retired batteries may experience alternating charge and discharge cycles during actual operation, directly calculating capacity using the cumulative ampere-hours and net state of charge (SBC) changes over the entire scheduling cycle would lead to an overestimation of capacity characteristics. Therefore, this invention selects effective sub-intervals where the SBC changes monotonically within each scheduling cycle for capacity characteristic calculation. Specifically, it selects continuous sub-intervals where the SBC either increases or decreases monotonically within the scheduling cycle, and where the absolute value of the SBC change is largest, as... The ratio of the time integral of the cumulative absolute value of the current within the sub-interval to the absolute value of the change in state of charge is used as the capacity retention characteristic.

[0031] If there is no valid monotonic sub-interval that meets the preset state of charge change threshold within the current scheduling cycle, the capacity maintenance characteristic value obtained from the most recent valid calculation will be used to ensure the continuity of the health carrying capacity index calculation process.

[0032] (2) Battery equivalent internal resistance characteristics : , in, Indicates the first The battery's equivalent internal resistance per scheduling cycle and These represent the voltage change and current change during the scheduling period, respectively. To prevent the default minimum positive number with a denominator of zero, the absolute value sign indicates that the absolute value of the corresponding change is taken.

[0033] (3) Battery charge and discharge efficiency characteristics : , in, Indicates the first Battery charge / discharge efficiency per scheduling cycle The scrolling statistics window corresponding to the efficiency calculation Indicates the battery discharge power. Indicates battery charging power. To prevent the use of a pre-defined minimum positive number with a denominator of zero.

[0034] Considering the randomness of charging and discharging conditions during the actual operation of retired batteries, the charging and discharging energy within a single scheduling cycle may not form a complete energy conversion process. Therefore, a rolling statistical window covering several recent scheduling cycles is used for calculation. The rolling statistical window selects a continuous time interval between the start and end times where the absolute value of the change in state of charge is less than a preset threshold. This ensures that the cumulative charged energy and cumulative released energy within the window correspond to approximately the same range of state of charge, thereby reducing the impact of differences in different state of charge intervals on the efficiency calculation results.

[0035] Furthermore, constraints are applied to the calculated charge / discharge efficiency characteristics: , in, This represents the lower limit of the efficiency characteristic, used to avoid distortion of the efficiency characteristic due to abnormal data; when there is no valid rolling statistical window that meets the conditions in the current scheduling period, the charge and discharge efficiency characteristic value obtained from the most recent valid calculation is used.

[0036] (4) Battery temperature rise rate characteristics : , in, Indicates the first Battery temperature rise rate per scheduling cycle and These represent the battery temperatures at the beginning and end of the scheduling cycle, respectively. This represents the scheduling cycle length.

[0037] (5) Battery voltage response deviation characteristics : , in, Indicates the first Battery voltage response deviation over one scheduling cycle and These represent the battery terminal voltage and operating current, respectively. This represents the open-circuit voltage corresponding to the current state of charge. This represents the battery's equivalent internal resistance. The reference voltage is indicated by the absolute value symbol, which represents the absolute value of the deviation from the terminal voltage response. The operating current is also included. By adopting the aforementioned convention of positive charging direction, a signed value is used so that the formula represents the deviation between the measured terminal voltage and the expected value of the equivalent circuit model under both charging and discharging conditions.

[0038] Based on the above five types of health degradation characteristics, a health degradation feature vector for retired batteries is constructed. : , in, For the first The health degradation feature vector of retired batteries for each scheduling cycle; , , , and These are the aforementioned capacity estimation, equivalent internal resistance, charge / discharge efficiency, temperature rise rate, and voltage response deviation characteristics, respectively. Indicates the rated capacity. This represents the initial equivalent internal resistance. This represents the reference temperature rise rate. A weighted fusion of multidimensional health degradation characteristics yields the health carrying capacity index of the retired battery. : , in, For the first The health carrying capacity index of retired batteries in each scheduling cycle. For health degradation feature vector The first in A dimensionless health degradation characteristic value For the first Non-negative weights of each health degradation characteristic Let be the health degradation feature index, 5 be the dimension of the health degradation feature, the exponential function represents the exponential operation with the natural constant as the base, and the summation symbol represents the summation of the weighted results of the five health degradation features. The weights satisfy: , in, ≥0 indicates that the weight of each health degradation feature is a non-negative number, and all... The sum equals 1, indicating that the five weights have been normalized. and The other meanings are the same as in the previous formula.

[0039] Feature weights The safe charge / discharge power limit is determined based on the comprehensive sensitivity of the corresponding health degradation characteristics in historical operating samples or bench test samples to the upper limit of safe charge / discharge power. The upper limit of safe charge / discharge power refers to the maximum charging and discharging power that a retired battery can safely perform under the corresponding health degradation characteristic values, as measured and calibrated in historical operating samples or bench test samples, rather than the values ​​calculated by the safe power boundary generation model of this invention, to avoid a circular dependency between weight calibration and boundary generation. While keeping other characteristics unchanged, the upper limit of safe charge / discharge power is determined based on the comprehensive sensitivity of the corresponding health degradation characteristics in historical operating samples or bench test samples to the upper limit of safe charge / discharge power. A preset small perturbation is applied to each health degradation characteristic. The rate of change of the maximum safe charging power and the maximum safe discharging power relative to the perturbation is calculated. The average of the absolute values ​​of the two rates of change is taken as the first... The sensitivity to each health degradation characteristic was calculated, and the five sensitivities were normalized to obtain... When sufficient historical samples are lacking, the initial weights calibrated by bench tests are used and updated based on subsequent actual power tracking deviations and battery temperature rise data.

[0040] The health carrying capacity index is directly input into the safe power boundary generation model, and the safe power boundary generation model outputs... and and as the first The upper and lower limits of the charging and discharging power of retired batteries in the feasible domain during each scheduling cycle.

[0041] The maximum safe discharge power boundary is: , in, For the first The maximum safe discharge power boundary for each scheduling cycle For the reason Determined discharge power health derating factor; minimum value operation is used to select the most stringent one from the candidate upper limits; This represents the maximum output power of the energy storage converter. This refers to the battery terminal voltage. This is the battery's maximum allowable current. This is the open-circuit voltage corresponding to the current state of charge. This is the lower limit of the battery's safe voltage. This is the equivalent internal resistance of the battery. To preset a very small positive number, This is the lower limit of the safe state of charge. For the battery's rated energy, The duration corresponding to the power boundary, where dis represents discharge and safe represents the safety boundary.

[0042] The maximum safe charging power boundary is: , in, For the first The maximum safe charging power boundary for each scheduling cycle For the reason The determined charging power healthy derating factor, This is the upper limit of the battery's safe voltage. This represents the upper limit of the safe state of charge, where ch indicates charging and safe indicates the safety boundary; the meanings of the other symbols are the same as those in the previous formula for the safe discharge power boundary.

[0043] , in, and They are respectively The discharge power health derating factor and the charging power health derating factor both increase monotonically. The health carrying capacity index of retired batteries, and These are the positive derating parameters for discharge and charge directions, calibrated through historical safe power samples or bench tests. After applying the health derating factor, Reducing it will directly narrow the executable power range; when When indicating a serious fault or isolation alarm, command = =0, causing the retired battery to be removed from the schedule.

[0044] Health degradation characteristics can be further modified by combining the cumulative number of cycles.

[0045] Furthermore, based on photovoltaic physics prediction, historical prediction errors and irradiance fluctuation characteristics are introduced to construct a photovoltaic power output reliability index and generate reliable photovoltaic available power that can be used for power balance constraints, as follows: First, the future [number]th ... Theoretical available photovoltaic power for each prediction step: , in, For the first The theoretical photovoltaic available power at the h-th prediction step within each scheduling cycle is used to select the smaller value between the inverter's rated power and the power calculated by the physical model. Indicates the rated power of the inverter. This represents the total area of ​​the photovoltaic modules. Indicates the reference conversion efficiency. Indicates predicted irradiance. Indicates the temperature coefficient. Indicates the predicted component temperature. Indicates reference temperature. This indicates a prediction of inverter efficiency. This indicates the prediction step size number.

[0046] By correcting the physical prediction results based on historical prediction errors, the predicted value of photovoltaic power output is obtained: , in, For the first Predicted value of photovoltaic output at the h-th prediction step within a scheduling cycle. To correspond to the theoretical available photovoltaic power of the prediction step size, This is the error correction amount obtained based on historical prediction errors.

[0047] Among them, historical prediction error for: , in, For the first Photovoltaic prediction error of a historical sample, For the first Actual photovoltaic output power of a historical sample For the photovoltaic power prediction corresponding to historical samples, This refers to the historical sample number.

[0048] Calculate the standard deviation of photovoltaic prediction error by selecting the most recent M historical scheduling cycles: , in, As of the date The standard deviation of the photovoltaic prediction error is calculated for each scheduling cycle, where M is the number of the most recent historical scheduling cycles involved in the calculation. The first one defined in the above formula Prediction error for each historical sample for The average of the prediction errors.

[0049] A photovoltaic power output prediction interval is constructed based on the standard deviation of the prediction error, with the lower and upper bounds as follows: , in, This is the lower bound of the photovoltaic power output prediction range. This is the predicted value for photovoltaic power output. The interval coefficients are determined based on the preset confidence level. The standard deviation of the photovoltaic prediction error is used, and the maximum value is taken to ensure that the lower bound of the prediction interval is not less than zero. Furthermore, The values ​​are as follows: , in, This represents the upper bound of the photovoltaic power output prediction range. The minimum value is taken as the inverter's rated power to ensure that the upper limit of the prediction interval does not exceed the inverter's rated power. , and The meaning is the same as in the previous formula.

[0050] To further integrate the prediction error level and the degree of irradiance variation, a photovoltaic prediction reliability index is constructed: , in, For the first The photovoltaic prediction confidence of the h-th prediction step within a scheduling cycle is obtained by using an exponential function to map the error and irradiance fluctuation to a range of 0 to 1. The standard deviation of photovoltaic prediction error. This is the predicted value for photovoltaic power output. and These represent the predicted step size and the irradiance for the current scheduling cycle, respectively. This is a scale parameter for irradiation changes. To prevent the use of a pre-defined minimum positive number with a denominator of zero.

[0051] When the prediction error is small and the irradiance change is stable Approaching 1; when the prediction error is large or the irradiance changes drastically, Approximately 0. Generate reliable usable photovoltaic power based on reliability: , in, For the first The reliable photovoltaic available power at the h-th prediction step within each scheduling cycle To improve the reliability of photovoltaic forecasts, This is the lower bound of the photovoltaic power output prediction range.

[0052] when When the level is high, the reliable available photovoltaic power is close to the predicted value of the photovoltaic point, enabling rolling dispatch to fully utilize the expected photovoltaic output; when At lower levels, the reliable available photovoltaic power shrinks towards the lower bound of the prediction range, causing the charging and discharging power of retired batteries and the switching power at grid connection points to be arranged according to a more conservative photovoltaic output.

[0053] Corresponding to the reliable available photovoltaic power, obtain the local load forecast power in the scheduling forecast time domain. And read the reference value of the switching power at the grid connection point. The net power demand for each forecast step is determined by the reliable available photovoltaic power, the predicted local load power, and the reference value of the grid connection point switching power.

[0054] For rolling prediction time domain Each prediction step within Define the set of scheduling variables: , in, For the first Within the scheduling cycle, the first A set of scheduling variables for each prediction step size; For the charging and discharging power of retired batteries, >0 indicates charging. <0 indicates discharge; For photovoltaic power generation limitation and ≥0; For power exchange at the grid connection point, >0 indicates that electricity is purchased from the power grid. <0 indicates power being fed back to the grid; h indicates the prediction step number.

[0055] for =1, 2, ..., The charging and discharging power of retired batteries meets the following requirements: , in, and The first Maximum safe discharge power boundary and maximum safe charge power boundary for each prediction step size The charging and discharging power of the retired battery corresponding to the predicted step size; The sign convention is the same as in the previous formula.

[0056] The state of charge is updated using the following formula: , in, and These represent the state of charge of retired batteries at the current prediction step and the next prediction step, respectively. and These are charging efficiency and discharging efficiency, respectively. The charge / discharge power of the retired battery is represented by the plus sign in the upper right corner of the square brackets, indicating the positive part of the function. The duration of a single prediction step. Rated energy for retired batteries.

[0057] , in, and These are the lower and upper safety limits for the state of charge of retired batteries, respectively. The state of charge of the retired battery for the next prediction step.

[0058] Battery temperature is predicted using the following formula: , in, and These are the battery body temperatures at the current prediction step and the next prediction step, respectively. The coefficient of temperature rise per unit charge / discharge power. For the charging and discharging power of retired batteries, The duration of a single prediction step. The heat dissipation coefficient between the battery and the environment. To correspond to the ambient temperature of the predicted step size, the absolute value sign indicates that the temperature rise term takes the charge and discharge power amplitude.

[0059] , in, For the next predicted step, the battery body temperature, This is the battery's maximum safe temperature. The meaning is the same as in the previous formula.

[0060] Simultaneously satisfying the power constraints of the energy storage converter: , in, The charge / discharge power of the retired battery at the h-th prediction step. This represents the maximum output power of the energy storage converter. The absolute value symbol indicates that the power amplitude in both the charging and discharging directions of the retired battery is constrained by this upper limit.

[0061] For each prediction step in the rolling prediction time domain, the following power balance relationship is established: , in, For reliable photovoltaic available power, To limit the power generation of photovoltaic power, For power exchange at the grid connection point, Forecast power for local load, The charging and discharging power of retired batteries.

[0062] In the above equation, the effective photovoltaic power supply is the reliable available photovoltaic power minus the photovoltaic curtailment power; the power consumed is generated when the retired battery is charging, and during discharge... The negative value results in a power supply. Therefore, the rolling dispatch model satisfies the power balance between power generation, energy storage, power consumption, and the grid at each prediction step.

[0063] The switching power at the grid connection point must meet the following upper and lower limits: , in, For the first Power exchanged at the grid connection point in the h-th prediction step within a scheduling cycle. The maximum allowable reverse power transmission at the grid connection point. This represents the maximum allowable power purchase capacity at the grid connection point. Grid indicates the grid connection point, out indicates reverse transmission, in indicates power purchase, and max indicates the maximum allowable power purchase capacity.

[0064] The switching power at the grid connection point further meets the following power change rate constraints: , When h = 1, Take the actual switching power of the grid connection point collected during the current scheduling cycle. That is, the power change rate constraint simultaneously restricts the power change between adjacent steps in the rolling prediction time domain and the power change of the grid connection point relative to the current actual operating point of the first prediction step.

[0065] in, The grid connection point switching power is for the current predicted step size. The power exchanged at the grid connection point for the previous prediction step size, This represents the upper limit of the allowable change in grid-connected power between adjacent prediction steps. The absolute value sign indicates the magnitude of the power difference between adjacent prediction steps.

[0066] The upper limits of allowed reverse power transmission, allowed power purchase, and power change rate at the grid connection point are determined based on the grid connection agreement, the capacity of the distribution transformer, or values ​​issued by the superior energy management system. When adjusting the power of retired batteries alone is insufficient to simultaneously meet power balance and grid connection constraints, the rolling dispatch model limits photovoltaic output by increasing the photovoltaic curtailment power.

[0067] For each prediction step, the degradation cost of the retired battery is: , in, For the degradation cost of the retired battery at the h-th prediction step, For the first The health carrying capacity index for each scheduling cycle. For the charging and discharging power of retired batteries, For the rated power of retired batteries, The duration of a single prediction step. This is the cost coefficient for exceeding temperature limits. The temperature of the battery itself. The positive sign in the upper right corner of the square brackets indicates the positive part of the function, representing a safe temperature threshold that does not incur additional degradation costs due to temperature.

[0068] The cost of switching charge / discharge power between adjacent prediction steps is: , in, The cost of switching charge / discharge power for the h-th prediction step. and These represent the charging and discharging power of the retired battery at the current prediction step and the previous prediction step, respectively. The absolute value sign indicates the magnitude of the difference between the charging and discharging power of adjacent prediction steps.

[0069] Construct the rolling scheduling objective function corresponding to the k-th scheduling cycle: , in, For the first The scheduling objective function value for each scheduling cycle. The duration of a single prediction step; To limit the power generation of photovoltaic power, For electricity purchase price, For power exchange at the grid connection point, The cost of battery degradation during retirement To compensate for the cost of switching charging and discharging power, To improve the reliability of photovoltaic forecasts, For the charging and discharging power of retired batteries, This is a reference value for the switching power at the grid connection point. to The non-negative weights of the six objective terms are listed in order.

[0070] The objective function has six terms: the first term is used to reduce photovoltaic power curtailment, the second term is used to reduce electricity purchase costs, the third and fourth terms are used to suppress the aging of retired batteries and frequent power switching, the fifth term is used to limit high-power operations when the reliability of photovoltaic predictions is low, and the sixth term is used to make the power exchanged at the grid connection point close to the reference value and reduce power disturbances to the distribution network.

[0071] In the objective function to The coefficients are determined jointly based on the dimensional normalization coefficients of each objective item and operational preferences. Determined based on photovoltaic power consumption targets and curtailment penalty standards, Determined based on time-of-use pricing and the weighting of electricity purchase costs. Determined based on battery replacement cost and lifespan degradation model, Determined based on the permissible charge / discharge switching frequency. Determined based on photovoltaic forecast error level and operational risk tolerance. The weights are determined based on the power point tracking requirements and allowable power fluctuations at the grid connection point as specified in the grid connection protocol. Each weight is first normalized according to the typical maximum value of the corresponding target item in historical samples, and then calibrated through historical dispatch playback or on-site trial operation.

[0072] In one implementation, the power step length is set according to a preset power step length. Discretize the charge / discharge power of retired batteries, the power curtailment of photovoltaic power, and the switching power at the grid connection point to obtain candidate power sequences; in the case of... Within the rolling prediction time domain of each prediction step, candidate power sequences that do not meet the following constraints are eliminated: safe charge / discharge power boundary of retired batteries, state of charge constraint, grid connection point switching power constraint, power change rate constraint, photovoltaic power curtailment constraint, and power balance constraint. The scheduling objective function value corresponding to the remaining candidate power sequences is then calculated. Select to make The smallest candidate power sequence is selected as the optimal scheduling power sequence.

[0073] The solution results of step S5 include the retired battery charge / discharge power sequence, the photovoltaic curtailment power sequence, and the grid connection point exchange power sequence, all of which satisfy the same power balance relationship and the same scheduling feasible region constraint. These three sequences together constitute the optimal scheduling power sequence, and none of them can be executed independently of the others.

[0074] The three optimal power sequences obtained in step S5 are represented as follows: , , , in, , and The first The optimal charging and discharging power sequence of retired batteries, the optimal photovoltaic power limiting sequence, and the optimal switching power sequence of grid connection points are obtained by solving the scheduling cycle; , and These are the first power values ​​of the three sequences; the asterisk indicates the optimal value obtained through optimization. The number of prediction steps included in the rolling prediction time domain.

[0075] Using rolling time-domain control, only , and The target values ​​for the pending charge / discharge power, pending photovoltaic curtailment power, and grid-connected point exchange power for the current scheduling cycle are used as the basis for calculation. The power values ​​for the remaining prediction steps are retained as subsequent prediction results and are not directly executed. Thus, step S5 completes the multi-step optimization and the first step extraction, while step S6 completes the safe projection, coordinated execution, and feedback loop of the first step power combination.

[0076] Before issuing the command, the charging / discharging power to be executed is projected onto the safe power range corresponding to the current measured state to obtain the charging / discharging power command for the retired battery: , in, For the first The final charging and discharging power command for the retired battery issued in each scheduling cycle; and These are the maximum safe charging power boundary and the maximum safe discharging power boundary for the current scheduling period, respectively. This represents the first optimal power value of the retired battery output in step S5; cmd indicates the command value. If after projection... and If the difference is different, it means that the current measured safety boundary has shrunk further compared to the boundary used in step S5. The photovoltaic power limiting value output in step S5 cannot be directly used. Otherwise, the power combination in the first time period will no longer meet the power balance.

[0077] To maintain the grid connection target for the first time period in step S5 and eliminate the power gap caused by the safety projection, the corresponding photovoltaic power limiting command is calculated based on the current reliable available photovoltaic power, the current local load power, the power command for charging and discharging retired batteries after projection, and the grid connection point exchange power target value output in step S5: , in, For the first The photovoltaic power curtailment instruction issued at the end of each scheduling cycle; The first prediction step size of the output from step S3 represents the reliable available photovoltaic power. The target power value for the first grid connection point output in step S5; The local load prediction power is the first prediction step size output from step S3; This is the command for the charging and discharging power of the retired batteries after projection. The maximum value is used to ensure the photovoltaic power curtailment is not less than 0, and the minimum value is used to ensure the photovoltaic power curtailment does not exceed the reliable available photovoltaic power. After calculation according to this formula... + = + This ensures that the battery power command and photovoltaic power curtailment command issued by S6 continue to maintain the power balance and grid connection target determined by S5 for the first time period. If the equation cannot be satisfied due to the upper and lower limits of photovoltaic power curtailment, the situation is determined to be scheduling infeasible, the scheduling feasibility domain is tightened according to the latest state, and S5 is re-executed.

[0078] When a serious fault or isolation alarm occurs in the battery management system, the above non-zero power combination will not be executed, and the battery will be immediately... Set to zero, and follow the updated... Recalculate .Will The command is sent to the energy storage converter, which then controls the charging or discharging of the retired battery according to the power command. This is converted into a photovoltaic output power limit command and sent to the photovoltaic inverter.

[0079] Within the current scheduling cycle, the actual charging and discharging power of the energy storage converter, the actual output power of the photovoltaic inverter, the actual state of charge of the retired battery, the actual temperature of the retired battery, the actual switching power at the grid connection point, and the alarm status of the battery management system are collected, and the power tracking deviation is calculated. , in, For the first Power tracking deviation of retired batteries in each scheduling cycle; This refers to the actual charging and discharging power of the energy storage converter. This refers to the charging and discharging power command issued for retired batteries; the absolute value symbol indicates the magnitude of the difference between the actual power and the commanded power; act represents the actual value, and cmd represents the commanded value.

[0080] Step S6 determines the operational anomaly level of the retired battery based on the power tracking deviation, the actual exchange power at the grid connection point, and alarm information from the battery management system, and handles it in the order of "first implementing safety protection, then maintaining grid connection hard constraints, and finally optimizing economic objectives." Specifically, operational anomalies of retired batteries are categorized as follows: Minor anomalies refer to power tracking deviations exceeding the Level 1 deviation threshold but not exceeding the Level 2 deviation threshold, or directional derating greater than 0 and less than 1; Moderate anomalies refer to power tracking deviations exceeding the Level 2 deviation threshold, actual power direction opposite to the command direction, energy storage converter failure, or actual exchange power at the grid connection point exceeding the allowable range and recoverable through rescheduling; Severe anomalies are triggered by any stop condition, Level 3 alarm, severe fault, or isolation alarm. Stop conditions refer to the battery's highest temperature, maximum temperature difference, or temperature rise rate reaching the corresponding stop value, causing the corresponding directional derating coefficient to be 0; Level 3 alarms are the highest level alarms defined by the battery management system; severe faults include thermal runaway, overcurrent tripping, high-voltage interlock disconnection, main contactor failure, insulation failure, and continuous communication interruption; isolation alarms refer to the battery management system's active isolation request. In the event of a minor anomaly, the charging and discharging power command for the current scheduling cycle is limited using the corrected safe charging and discharging power range, and the photovoltaic power limiting command is also corrected accordingly. In the event of a moderate anomaly, the actual charging and discharging power, the actual exchange power at the grid connection point, and the latest measured status are written into the initial status of the next scheduling cycle. S2 to S5 are executed based on the latest measured status to update the health carrying capacity index, safe charging and discharging power boundary, scheduling feasible domain, and scheduling power command. In the event of a severe anomaly, charging or discharging in the affected direction is prohibited. If both the charging and discharging directions are prohibited, the charging and discharging power command for retired batteries is set to zero, and the photovoltaic power limiting command is recalculated accordingly.

[0081] When any of the following abnormal conditions are triggered: power tracking deviation, grid connection power, temperature, state of charge, or battery management system alarm, first update the health carrying capacity index, safe charge / discharge power boundary, and scheduling feasible region based on the latest measured status; for stop thresholds or serious faults, immediately... Set to zero and recalculate the corresponding components. For situations where temperature, state of charge, or directional alarms enter the derating range, the safe power boundary in the corresponding direction is first narrowed, and then S5 is re-executed. For situations where power tracking is abnormal or the grid connection point power exceeds the limit, the actual power and the measured power at the grid connection point are written into the new initial state and constraints, and then S5 is re-executed and the corrected first-period power combination is issued. Therefore, S6 is not an additional protection independent of S5, but rather it converts the first-period optimization result of S5 into executable instructions and feeds back the execution deviation as the state variables and boundary conditions for the next S5 solution.

[0082] To ensure a one-to-one correspondence between the triggering conditions, scheduling objects, and handling actions in step S6, the highest temperature of the retired battery is defined as... The maximum temperature difference is defined as The rate of temperature rise is defined as The actual state of charge is defined as The alarm level of the battery management system is defined as follows: The above quantities, together with the power tracking deviation and the actual switching power at the grid connection point, constitute the feedback judgment quantity of S6.

[0083] Temperature thresholds are represented by warning and stop values, and must meet the following conditions: < , < , < as well as < The state of charge threshold is satisfied. < < < Power tracking deviation uses a first-level deviation threshold. and second-order deviation threshold ,and < Each threshold is pre-calibrated based on the battery chemistry system, manufacturer protection values, degree of decommissioning, and energy storage converter capacity, and should not be relaxed during the scheduling process due to optimization objectives.

[0084] The maximum temperature, maximum temperature difference, and rate of temperature rise each form an independent derating coefficient; minimum and maximum values ​​are used to limit each coefficient to between 0 and 1. , , , , , in, Indicates the first Within each scheduling cycle, the charging direction is determined by the highest temperature of the retired battery, and the maximum temperature reduction factor is used to determine the charging direction. Indicates the first Within a scheduling cycle, the discharge direction is determined by the highest temperature of the retired battery, and the highest temperature depreciation coefficient is used to determine the discharge rate. Indicates the first The maximum temperature difference reduction coefficient is determined by the maximum temperature difference of the retired battery within each scheduling cycle; Indicates the first The depreciation coefficient of the temperature rise rate determined by the temperature rise rate of the retired battery within each scheduling cycle; and They represent the first The charging direction temperature derating coefficient and the discharging direction temperature derating coefficient are obtained by combining the highest temperature, the maximum temperature difference, and the temperature rise rate within a scheduling cycle. and These represent the maximum temperature stop thresholds for the charging and discharging directions, respectively. and These represent the maximum temperature warning thresholds for the charging and discharging directions, respectively. Indicates the first The highest measured temperature among all effective temperature measurement points of retired batteries within a scheduling cycle; and These represent the maximum temperature difference stop threshold and the maximum temperature difference warning threshold, respectively. Indicates the first The difference between the highest and lowest measured temperatures at each effective temperature measurement point of the retired battery within a scheduling cycle; and These represent the temperature rise rate stopping threshold and the temperature rise rate warning threshold, respectively. Indicates the first The actual temperature rise rate of retired batteries within a scheduling cycle.

[0085] The maximum temperature sub-coefficient applies only to the corresponding charging or discharging direction, while the maximum temperature difference sub-coefficient and the temperature rise rate sub-coefficient apply to both directions simultaneously; when either index reaches the stopping threshold, the corresponding derating factor is set to 0. State of charge derating is only applied directionally. , , in, This represents the derating factor of the charging direction, determined by the actual state of charge of the retired battery. This represents the state-of-charge derating factor, determined by the actual state of charge of the retired battery in the discharge direction. and These represent the upper and lower limits of the safe state of charge for retired batteries, respectively, and are used as the thresholds for stopping charging and stopping discharging. Indicates the first The actual state of charge of retired batteries within a scheduling cycle; High charge state warning threshold indicating the charging direction; The low state of charge warning threshold indicates the direction of discharge. All state of charge parameters use the same metering caliber.

[0086] Battery management system alarms are only used to determine the alarm derating factor and whether the system has entered a protection state; a separate scheduling process is no longer set up for each type of alarm. When the alarm level is 0, the alarm derating factor for both the charging and discharging directions is set to 1. When it is Level 1 or Level 2, only the alarm rate reduction factor in the affected direction will be applied. or Set to a pre-calibrated value between 0 and 1 that decreases with the alarm level; When the alarm derating factor is set to 0 in both directions, it is level 3, or when any of the following serious faults occur: thermal runaway, overcurrent tripping, high-voltage interlock disconnection, main contactor failure, insulation failure, battery management system active isolation request, or continuous communication interruption.

[0087] Temperature, state of charge, and alarm derating are applied uniformly to the safe power boundary obtained in step S2 to obtain the corrected boundary actually used in S6: , , In an embodiment of the present invention, step S6 processes all feedback states according to the following unified rules: when the power tracking deviation is not greater than the first-level deviation threshold, the actual switching power at the grid connection point does not exceed the limit, and all derating factors are 1, the actual state of charge, actual temperature, actual charging and discharging power, and actual switching power at the grid connection point are taken as the initial state for the next scheduling cycle; when the power tracking deviation exceeds the first-level deviation threshold but does not exceed the second-level deviation threshold, or any derating factor is greater than 0 and less than 1, it is determined to be a minor anomaly, and the interval [ , The charging and discharging power commands for the current scheduling cycle are limited, and the photovoltaic power limiting command is recalculated simultaneously. When the power tracking deviation exceeds the second-level deviation threshold, the actual power direction is opposite to the command direction, the energy storage converter fails, or the actual exchange power of the grid connection point exceeds the limit and can be restored through rescheduling, it is determined to be a medium-level anomaly. The actual power, the measured power of the grid connection point, and the latest corrected boundary are written into the initial state and constraints of the next scheduling cycle. S2 to S5 are executed with the latest measured state to update the health carrying capacity index, safe charging and discharging power boundary, scheduling feasible domain, and scheduling power command. When any stop condition, level 3 alarm, serious fault, or isolation alarm causes the derating factor of the corresponding direction to be 0, it is determined to be a serious anomaly. The retired battery is removed from the current charging and discharging scheduling cycle, and charging or discharging in the corresponding direction is prohibited. When both charging and discharging directions are prohibited, the executable interval is set to [0, 0] and the charging and discharging power command of the retired battery is set to zero.

[0088] For grid-connected point backfeed power that cannot be eliminated within the safe charging and discharging power boundary of retired batteries, photovoltaic power limiting is increased first; each time the charging and discharging power command of retired batteries changes due to limiting or zeroing, the photovoltaic power limiting command is recalculated synchronously according to the power balance relationship in step S6, so as to maintain the grid-connected point exchange power target determined in S5.

[0089] The release of limit or zeroing adopts a unified recovery condition: the non-zero power range in the corresponding direction is only restored after the triggering factor disappears and the relevant feedback quantity continuously meets the preset recovery condition; serious faults or isolation alarms also require a grid connection permission signal from the battery management system. After recovery, S5 re-optimizes the power combination of the first time period, without directly restoring the instructions before the anomaly. Thus, all anomalies are categorized into three types of scheduling actions: "state recursion, re-solving after boundary contraction, and linkage recalculation after protection zeroing".

[0090] For example, within a representative dispatch cycle, the credible available photovoltaic power is 60kW, the predicted local load power is 35kW, and the photovoltaic power is 25kW surplus relative to the local load. This is based on the current retired battery health carrying capacity index. =0.62 yields a safe charging power boundary of 18kW, and the maximum allowable reverse power transmission at the grid connection point is 5kW. The power balance relationship for this scheduling cycle is: , in, To limit the power generation of photovoltaic power, For power exchange at the grid connection point, For the charging and discharging power of retired batteries, This represents the current dispatch cycle number; a negative grid connection power indicates power is being fed back to the grid, while a positive charge / discharge power of a retired battery indicates charging.

[0091] The S5 received an initial charging power of 18kW for the retired battery and a grid connection power of [missing information]. With a power output of 5kW and a photovoltaic (PV) power limit of 2kW, and the S6 projection boundary not shrinking, a charging power command of 18kW and a PV power limit command of 2kW are issued. If the actual charging power of the energy storage converter is 17.4kW and the power tracking deviation does not exceed the first-level deviation threshold, and the other feedback quantities are normal, then the actual state will be used for the next scheduling cycle. This result simultaneously satisfies the healthy power boundary, the grid connection point backfeed power constraint, and power balance.

[0092] Under the same operating conditions described above, if the latest feedback causes the charging direction correction boundary to shrink to 10.8kW, then the S6 will limit the battery power command to 10.8kW and follow the same... The 5kW grid-connected target will recalculate the photovoltaic power limiting to 9.2kW. If a stop condition or serious alarm occurs, causing the correction boundaries in both the charging and discharging directions to be 0, the battery power command will be set to zero, and the photovoltaic power limiting will be recalculated to 20kW. This same calculation example illustrates the normal initial execution of S5, the safety limiting of S6, and the photovoltaic coordinated compensation after setting the power to zero.

[0093] Under the same operating conditions described above, if the fixed power boundary is set at 25kW, it will exceed the current health and safety boundary of 18kW; if it is fixed at 12.5kW, it will increase photovoltaic power limitation when the battery can still safely absorb 18kW. In contrast, this invention forms a dynamic safety boundary in S2, obtains the optimal initial power combination of battery-photovoltaic-grid connection point within this boundary in S5, and then performs safety projection and linkage recalculation based on the latest measured state in S6, reflecting the progressive relationship between health status, rolling optimization and closed-loop execution.

[0094] Example 2: This invention provides a closed-loop scheduling system for the on-site consumption of decommissioned photovoltaic cells based on dynamic boundaries of health carrying capacity. For example... Figure 2 As shown, the system includes a data sensing module 201, a health boundary generation module 202, a reliable prediction and power balance module 203, a rolling scheduling optimization module 204, a control execution module 205, and a feedback correction module 206.

[0095] The data sensing module 201 is used to collect operating data of retired batteries, photovoltaic power generation, local load, grid connection points and power conversion equipment according to the scheduling cycle, and to perform time alignment, outlier processing and filtering on data from different sources.

[0096] The health boundary generation module 202 is used to extract capacity degradation, equivalent internal resistance, energy efficiency, temperature rise rate and voltage response deviation characteristics from the operating data of retired batteries, calculate the health carrying capacity index, and generate safe charging power boundary and safe discharging power boundary based on the health carrying capacity index and voltage, current, state of charge, temperature and energy storage converter limits.

[0097] The reliable prediction and power balance module 203 is used to generate reliable photovoltaic available power and local load predicted power in the rolling prediction time domain, read grid connection point power limits and reference values, and establish a power balance relationship between reliable photovoltaic available power, photovoltaic power limit, grid connection point exchange power, local load power and retired battery charging and discharging power.

[0098] The rolling scheduling optimization module 204 is used to construct the scheduling feasible region based on the healthy power boundary, battery state constraints, energy storage converter power constraints, power balance relationship and grid connection point safety constraints, and solve the rolling scheduling objective function, outputting the retired battery charging and discharging power sequence, photovoltaic power limiting power sequence and grid connection point exchange power sequence that mutually satisfy the same power balance relationship.

[0099] The control execution module 205 is used to extract three power sequences: the retired battery charging and discharging power sequence, the photovoltaic power limiting power sequence, and the grid connection point switching power sequence. The first power value of each sequence is projected onto the latest corrected safe power range. Based on the projected retired battery power command, the first power target of the grid connection point, the reliable photovoltaic available power, and the local load power, the photovoltaic power limiting power command is calculated and then sent to the energy storage converter and the photovoltaic inverter for execution.

[0100] The feedback correction module 206 is used to obtain the actual charging and discharging power of the energy storage converter, the actual output power of the photovoltaic inverter, the actual state of charge and temperature of the retired battery, the actual switching power at the grid connection point, and the alarm status of the battery management system. When the feedback status is normal, the actual status is passed to the next scheduling cycle. When the derating condition is triggered, the safety power boundary in the corresponding direction is shrunk and the rolling scheduling optimization module 204 is recalled. When the stop condition, serious fault, or isolation alarm is triggered, the power command of the retired battery is set to zero and the photovoltaic power limiting command is recalculated synchronously.

[0101] The specific implementation process of each module, as well as the related conditions and concepts, are described in the corresponding method description in Example 1. To avoid redundancy, they will not be described in detail here.

[0102] Example 3: like Figure 3As shown, the present invention provides an electronic device 300, including a processor 301, a communication interface 302, a memory 303, an input / output interface 304, and a bus 306; the processor 301, communication interface 302, memory 303, and input / output interface 304 are respectively connected to the bus 306. This embodiment also provides a computer program product, including a computer program or instructions; when the computer program or instructions are executed by the processor 301, they implement… Figure 1 The steps of the closed-loop scheduling method for on-site consumption of decommissioned photovoltaic cells based on dynamic boundary of health carrying capacity are shown.

[0103] The memory 303 serves as a computer-readable storage medium for storing scheduler instructions 305; the processor 301 reads and executes the scheduler instructions 305 to implement... Figure 1 The method for dispatching decommissioned batteries for local photovoltaic consumption is shown. Communication interface 302 is used for data communication with the decommissioned battery management system, energy storage converter, photovoltaic inverter and grid connection point metering device; input / output interface 304 is used to receive operating parameters and output dispatch results and power control commands.

[0104] This embodiment also provides a computer-readable storage medium storing scheduler instructions 305 thereon; when the scheduler instructions 305 are executed by the processor 301, they implement... Figure 1 This paper presents a closed-loop scheduling method for the local consumption of retired photovoltaic cells.

[0105] In summary, this invention directly generates safe charging and discharging power boundaries from the health carrying capacity index and uses these boundaries to define the feasible domain for rolling scheduling. Power allocation among photovoltaics, local loads, retired batteries, and the grid is determined through explicit power balance relationships and grid connection point safety constraints, forming charging and discharging power commands and photovoltaic power limiting commands that can be issued to the energy storage converter and photovoltaic inverter, respectively. The next scheduling cycle is then recalculated based on actual power, battery status, grid connection point exchange power, and battery management system alarms, thereby achieving closed-loop coordination between the health status of retired batteries, local photovoltaic consumption demand, grid-connected operation safety, and power control execution.

[0106] Those skilled in the art will understand that the methods disclosed in the above embodiments can be implemented by software plus necessary hardware. Based on this understanding, the technical solution of this application, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions to cause a computer device to execute the methods described in the above embodiments.

Claims

1. A closed-loop scheduling method for on-site consumption of decommissioned photovoltaic power generation batteries, characterized in that, The following steps are executed according to a continuous scheduling cycle: S1: Collect operational data from the retired battery side, photovoltaic power generation side, local load side, grid connection point, and power conversion equipment side; among which, the operational data from the retired battery side includes battery terminal voltage, operating current, state of charge, battery body temperature, and battery management system alarm information; the operational data from the photovoltaic power generation side includes actual photovoltaic output power, and at least one of solar irradiance, photovoltaic module temperature, and weather forecast data; the operational data from the local load side includes local load power; and the operational data from the grid connection point and power conversion equipment side includes grid connection point exchange power, maximum allowable power purchase capacity, maximum allowable reverse power supply capacity, maximum grid connection point power change rate, and maximum output power of the energy storage converter. S2: Extract multidimensional health degradation features from the operating data of retired batteries, calculate the health carrying capacity index of retired batteries, and generate the safe charging power boundary and safe discharging power boundary for the current scheduling cycle based on the health carrying capacity index of retired batteries. S3: Based on the photovoltaic physical model, historical photovoltaic prediction errors, and the degree of irradiance variation, obtain the reliable photovoltaic available power in the scheduling prediction time domain, and obtain the local load prediction power in the scheduling prediction time domain; S4: Using the charging and discharging power of retired batteries, the curtailed power of photovoltaics, and the exchange power at the grid connection point as scheduling variables, establish the real-time power balance relationship between the generation-storage-consumption and the grid among the reliable available photovoltaic power, the curtailed power, the exchange power at the grid connection point, the local load power, and the charging and discharging power of retired batteries. Based on the safe charging power boundary, safe discharging power boundary, state of charge constraint, temperature constraint, power constraint of energy storage converter, and grid security constraint, a scheduling feasible domain is constructed. The grid security constraint includes the exchange power constraint at the grid connection point and the power change rate constraint at the grid connection point. S5: Within the feasible scheduling domain, with the optimization objectives of photovoltaic curtailment power generation, grid power purchase, degradation cost of retired batteries, charging and discharging power switching cost, photovoltaic prediction uncertainty risk and grid connection point power fluctuation, the retired battery charging and discharging power sequence and photovoltaic curtailment power sequence in the rolling optimization time domain are obtained. The rolling optimization time domain consists of multiple prediction steps, and the power value corresponding to the first prediction step in the two sequences is selected as the charging and discharging power command and photovoltaic curtailment power command for the current scheduling cycle, respectively. S6: The energy storage converter receives the charging and discharging power command for the current scheduling cycle, and the photovoltaic inverter receives the actual charging and discharging power of the retired battery. When the deviation between the actual charging and discharging power of the retired battery and the charging and discharging power command for the current scheduling cycle exceeds a preset deviation threshold, or the grid connection point exchange power exceeds the allowable range, or the battery management system generates an alarm, the abnormal operation level of the retired battery is determined based on the deviation between the actual charging and discharging power of the retired battery and the charging and discharging power command for the current scheduling cycle, or the grid connection point exchange power, or the alarm information generated by the battery management system. The abnormality level includes minor abnormality, moderate abnormality, or severe abnormality. When the retired battery is operating under a minor abnormality, the charging and discharging power command for the current scheduling cycle is limited and corrected. When the retired battery is operating under a moderate abnormality, S2 to S5 are executed based on the latest measured state, updating the health carrying capacity index, safe charging and discharging power boundary, and scheduling power command of the retired battery, and entering the next scheduling cycle. When the retired battery is operating under a severe abnormality, the retired battery exits the current charging and discharging scheduling cycle and the charging and discharging power command for the current scheduling cycle is set to zero.

2. The method according to claim 1, characterized in that, In step S2, the multidimensional health degradation features include capacity degradation features, equivalent internal resistance features, charge / discharge efficiency features, temperature rise rate features, and voltage response deviation features, and a health degradation feature vector for retired batteries is constructed. Health carrying capacity index of retired batteries for: in, Let k be the health carrying capacity index of the retired battery in the k-th scheduling cycle, where k is a natural number. Let be the health degradation feature vector of the retired battery in the k-th scheduling cycle; exp(·) is an exponential function with the natural constant as the base; k is the scheduling cycle number; i is the health degradation feature number, i = 1, 2, ..., 5; for The i-th dimensionless health degradation feature value in the data; The non-negative weight of the i-th health degradation feature; The weights satisfy: Among them, each A sum equal to 1 indicates that the weights have been normalized; 0 < ≤1, The smaller the value, the smaller the safe charging power boundary and safe discharging power boundary generated in step S2; The sensitivity of the i-th health degradation feature to the maximum safe charging power and maximum safe discharging power of the retired battery is determined by: based on historical operating samples or bench test samples, a preset small perturbation is applied to the i-th health degradation feature while keeping the other health degradation features unchanged. The rate of change of the maximum safe charging power and the maximum safe discharging power relative to the perturbation is calculated, and the average of the absolute values ​​of the two rates of change is taken as the sensitivity of the i-th health degradation feature. The five sensitivities are then normalized to obtain... When historical samples are insufficient, the initial weights calibrated by bench tests are used and updated based on actual power tracking deviations and battery temperature rise feedback.

3. The method according to claim 1, characterized in that, In step S2, the safe discharge power boundary for the current scheduling period. and safe charging power boundary They are respectively: in, This represents the maximum safe discharge power boundary for the k-th scheduling cycle; For the reason Determined discharge power health derating factor; This represents the maximum output power of the energy storage converter. The battery terminal voltage during the k-th scheduling cycle; This is the battery's maximum allowable current. To match the current state of charge The corresponding battery open-circuit voltage; This is the lower limit of the battery's safe voltage. ε is the equivalent internal resistance of the battery in the k-th scheduling cycle; ε is a preset positive number to prevent the denominator from being zero. This is the lower limit of the safe state of charge. τ represents the battery's rated energy; τ is the duration corresponding to the power boundary. in, This represents the maximum safe charging power boundary for the k-th scheduling cycle. For the reason Determined charging power healthy derating factor; This represents the maximum output power of the energy storage converter. The battery terminal voltage during the k-th scheduling cycle; This is the battery's maximum allowable current. This is the upper limit of the battery's safe voltage. To match the current state of charge The corresponding battery open-circuit voltage; Let ε be the equivalent internal resistance of the battery in the k-th scheduling cycle; ε is a preset positive number. This represents the upper limit of safety for the state of charge. τ represents the battery's rated energy; τ is the duration corresponding to the power boundary. in, and They are respectively The discharge power health derating factor and the charging power health derating factor both increase monotonically. The health carrying capacity index of the retired battery in the k-th scheduling cycle; and These are the positive derating parameters for the discharge and charging directions, respectively, calibrated from historical safe power samples or bench tests; dis represents discharge, and ch represents charging; when the battery management system issues a Level 3 alarm, a critical fault, or an isolation alarm... and All were set to zero.

4. The method according to claim 1, characterized in that, In step S3, the reliable photovoltaic prediction result for the h-th prediction step in the scheduling prediction time domain includes the predicted value of the photovoltaic point, the lower bound of the photovoltaic prediction interval, the upper bound of the photovoltaic prediction interval, the reliability of the photovoltaic prediction, and the reliable available photovoltaic power; the reliable available photovoltaic power is: in, The reliable available photovoltaic power at the h-th prediction step within the k-th scheduling period; To determine the reliability of photovoltaic predictions corresponding to the prediction step size, ∈[0,1]; Predict power for photovoltaic points; The lower bound of the photovoltaic prediction interval is given by k; the current scheduling cycle number is given by h; the prediction step number is given by pv; and the photovoltaic power is represented by pv. Step S3 also obtains the predicted local load power corresponding to each prediction step. This together forms the power balance relationship in step S4.

5. The method according to claim 1, characterized in that, In step S4, for each prediction step in the scheduling prediction time domain, the power balance relationship is established as follows: in, Reliable photovoltaic power; For photovoltaic power generation limitation and ≥0; For power exchange at the grid connection point, >0 indicates that electricity is purchased from the power grid. <0 indicates that power is being fed back into the grid; Predict power for local load; For the charging and discharging power of retired batteries, >0 indicates that the retired battery is being charged. <0 indicates the discharge of retired batteries; k is the current scheduling cycle number; h is the prediction step size number; pv, curt, grid, L, and b represent photovoltaic, curtailment, grid connection point, load, and battery, respectively; the grid connection point exchange power satisfies: in, The switching power of the grid connection point at the h-th prediction step within the k-th scheduling cycle; The maximum allowable reverse power transmission at the grid connection point; This represents the maximum allowable power purchase capacity at the grid connection point; grid indicates the grid connection point, out indicates reverse transmission, in indicates power purchase, and max indicates the maximum capacity. in, Power exchanged at the grid connection point for the current predicted step size; The grid-connected power exchanged at the previous prediction step; |·| is the absolute value operation; The maximum allowable change in grid-connected power between adjacent prediction steps; h is the prediction step number; grid represents the grid-connected point, and max represents the maximum allowable change.

6. The method according to claim 1, characterized in that, In step S5, in the case of including Within the time domain of scheduling prediction for each prediction step, the following scheduling objective function is constructed: in, is the scheduling objective function value for the k-th scheduling cycle; h is the prediction step number; The number of prediction steps included in the time domain of the scheduling prediction; Δt is the duration of a single prediction step; Limiting photovoltaic power generation; The electricity purchase price corresponding to the predicted step size; Power exchange at the grid connection point; It is a positive part function; The cost of degraded batteries; The cost of switching charging and discharging power between adjacent prediction steps; To assess the reliability of photovoltaic forecasts; The charging and discharging power of retired batteries; This is a reference value for the switching power at the grid connection point; to These are the non-negative weights for the photovoltaic power generation restriction item, electricity purchase cost item, battery degradation item, power switching item, photovoltaic forecast uncertainty risk item, and grid connection point power deviation item; the superscript ref indicates a reference value; to The following parameters are determined based on the dimensional normalization coefficients of each objective item and operational preferences: Determined based on photovoltaic power consumption targets and curtailment penalty standards, Determined based on time-of-use pricing and the weighting of electricity purchase costs. Determined based on battery replacement cost and lifespan degradation model. Determined based on the permissible charge / discharge switching frequency. Determined based on photovoltaic forecast error level and operational risk tolerance. Determined based on the power point tracking requirements and allowable power fluctuations at the grid connection point as specified in the grid connection protocol; within the dispatchable feasible region, the power point is determined. The minimum power sequence for retired batteries to charge and discharge and the power sequence for photovoltaic power limiting are used, and only the first power value in the two sequences is executed using a rolling time-domain method.

7. The method according to claim 1, characterized in that, Step S6 includes: projecting the charge / discharge power command of the current scheduling period onto the interval [- , The command is then sent to the energy storage converter, and the photovoltaic power limiting command for the current scheduling cycle is sent to the photovoltaic inverter; the actual output power of the energy storage converter is obtained. And the actual output power of the photovoltaic inverter, and calculate The deviation between the power output and the charge / discharge command is considered. If the deviation exceeds the first-level deviation threshold but does not exceed the second-level deviation threshold, or if the derating factor corresponding to temperature, state of charge, or alarm is greater than 0 and less than 1, it is determined to be a minor anomaly. The power output of the retired battery is limited using the corrected safe power range, and the photovoltaic power limiting command is recalculated. If the deviation exceeds the second-level deviation threshold, the actual output power direction is opposite to the command direction, the energy storage converter fails, or the grid connection point's switching power exceeds the allowable range and can be recovered through rescheduling, it is determined to be a moderate anomaly. S2 to S5 are executed based on the latest measured state. If any stop condition, a third-level alarm in the battery management system, a serious fault, or an isolation alarm is triggered, it is determined to be a serious anomaly. The power output of the retired battery is reset to zero, and the photovoltaic power limiting command is recalculated. Otherwise, the actual output power, the executed state of charge, and the temperature are used as the initial state for the next scheduling cycle. in,[- , [ ] represents the range of charging and discharging power allowed to be issued for retired batteries in the k-th scheduling cycle; This is the maximum safe discharge power boundary, and the preceding negative sign indicates that the discharge power is a negative value. This represents the maximum safe charging power boundary. This represents the actual output power of the energy storage converter in the kth scheduling cycle; the superscript 'safe' indicates the safety boundary, the superscript 'act' indicates the actual value, the subscript 'b' indicates the retired battery, the subscript 'dis' indicates discharging, and the subscript 'ch' indicates charging.

8. A closed-loop scheduling system for on-site consumption of decommissioned photovoltaic power generation batteries, characterized in that, include: The data sensing module is used to collect operating data of retired batteries, photovoltaic power generation, local load, grid connection points and power conversion equipment according to the scheduling cycle, and to perform time alignment, outlier processing and filtering on data from different sources; The health boundary generation module is used to extract capacity degradation, equivalent internal resistance, energy efficiency, temperature rise rate and voltage response deviation characteristics from the operating data of retired batteries, calculate the health carrying capacity index, and generate safe charging power boundary and safe discharging power boundary based on the health carrying capacity index and voltage, current, state of charge, temperature and energy storage converter limits. The reliable prediction and power balance module is used to generate reliable photovoltaic available power and local load predicted power in the rolling prediction time domain, read grid connection point power limits and reference values, and establish a power balance relationship between reliable photovoltaic available power, photovoltaic power curtailment, grid connection point switching power, local load power and retired battery charging and discharging power. The rolling scheduling optimization module is used to construct the scheduling feasible region based on the health power boundary, battery state constraints, energy storage converter power constraints, power balance relationship and grid connection point safety constraints, and solve the rolling scheduling objective function to output the retired battery charge and discharge power sequence, photovoltaic power curtailment power sequence and grid connection point exchange power sequence that mutually satisfy the same power balance relationship. The control execution module is used to extract three power sequences: the retired battery charging and discharging power sequence, the photovoltaic power curtailment power sequence, and the grid connection point switching power sequence. The first power value of each sequence is projected onto the latest corrected safe power range. Based on the projected retired battery power command, the first power target of the grid connection point, the reliable photovoltaic available power, and the local load power, the photovoltaic power curtailment command is calculated and then sent to the energy storage converter and the photovoltaic inverter for execution. The feedback correction module is used to obtain the actual charging and discharging power of the energy storage converter, the actual output power of the photovoltaic inverter, the actual state of charge and temperature of the retired battery, the actual switching power at the grid connection point, and the alarm status of the battery management system. When the feedback status is normal, the actual status is passed to the next scheduling cycle. When the derating condition is triggered, the safety power boundary in the corresponding direction is shrunk and the rolling scheduling optimization module is called again. When the stop condition is triggered or a serious fault or isolation alarm occurs, the power command of the retired battery is set to zero and the photovoltaic power curtailment command is recalculated synchronously.

9. A computer-readable storage medium storing a computer program, wherein when executed by a processor, the computer program implements the closed-loop scheduling method for on-site consumption of decommissioned photovoltaic cells as described in any one of claims 1 to 7.

10. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the closed-loop scheduling method for on-site consumption of decommissioned photovoltaic cells as described in any one of claims 1 to 7.

Images