Multi-stage dynamic scheduling method and system of photovoltaic energy storage street lamp at urban road intersection

Abstract

This application discloses a multi-level dynamic scheduling method and system for photovoltaic energy storage streetlights at urban road intersections, relating to the field of intelligent traffic lighting control technology. The disclosed multi-level dynamic scheduling method and system for photovoltaic energy storage streetlights at urban road intersections acquires traffic phase time-series data and real-time traffic flow density data at the intersection, performs fusion processing to generate predicted lighting load power curve data, identifies high-power pulse load segments, controls the charging and discharging state of energy storage units, and dynamically allocates power during high-power periods to meet real-time needs. It can proactively adapt to changes in traffic flow, finely coordinate the operation of energy storage units, effectively extend the lifespan of energy storage units, and improve lighting stability.

Description

Technical Field

[0001] This application relates to the field of intelligent traffic lighting control technology, and in particular to a method and system for multi-level dynamic scheduling of photovoltaic energy storage streetlights at urban road intersections. Background Technology

[0002] With the increasing density of urban transportation networks and the deepening of intelligent management, photovoltaic energy storage streetlights have been widely adopted in road lighting infrastructure due to their advantages of low carbon emissions, environmental friendliness, and deployment flexibility. However, in the highly dynamic urban road intersection scenario, such systems face unique operational challenges. The lighting load at intersections is directly affected by the periodic switching of traffic lights and vehicle traffic behavior, exhibiting instantaneous and drastic fluctuation characteristics. For example, at the instant the light turns from red to green, a large number of vehicles start simultaneously, causing the lighting power demand to surge to its peak within seconds; or during the transition from green to red, vehicles decelerate and stop, causing the power demand to drop rapidly. This power pulse characteristic on the order of seconds or even sub-seconds poses a severe challenge to existing photovoltaic energy storage systems.

[0003] Current mainstream technologies generally adopt a linear architecture of "photovoltaic panel-battery-LED light," whose energy management strategies rely on historical average load data for static design, lacking the ability to perceive and respond to real-time traffic dynamics. When encountering high-power pulse loads unique to intersections, the system cannot adjust the operating state of the energy storage unit in time, forcing the battery to withstand a large current discharge impact in a short period of time. This operating mode exacerbates the electrochemical polarization phenomenon inside the battery, not only leading to accelerated battery capacity decay and shortened cycle life, but also potentially causing local heat accumulation and increasing the risk of thermal runaway. At the same time, due to the limited instantaneous output capability of the battery, the LED lighting module has difficulty maintaining stable light output, and is prone to brightness fluctuations or flickering, which can easily interfere with the driver's vision at night or in low visibility conditions, posing a potential threat to traffic safety.

[0004] While some improvements attempt to address load fluctuations by increasing battery capacity or boosting photovoltaic panel rated power, these measures only address hardware redundancy and fail to optimize the system's dynamic response capabilities at the fundamental level of energy dispatch mechanisms. The result is increased system costs and space requirements, yet the damage to energy storage units caused by pulsed loads remains unresolved, severely limiting the long-term reliable operation and economic feasibility of photovoltaic energy storage streetlights in highly dynamic traffic scenarios.

[0005] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0006] The main purpose of this application is to provide a multi-level dynamic scheduling method and system for photovoltaic energy storage streetlights at urban road intersections, which aims to extend the life of energy storage units and improve lighting stability.

[0007] To achieve the above objectives, this application proposes a multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections, the method comprising: Acquire traffic phase time-series data and real-time traffic flow density data at the intersection, and perform fusion processing on the traffic phase time-series data and the real-time traffic flow density data to generate predicted lighting load power curve data; Based on the predicted lighting load power curve data, high-power pulse load segment data is identified, and pre-scheduling instruction data is generated according to the high-power pulse load segment data. Based on the pre-scheduling instruction data, the charging and discharging state of the high-power-density buffer unit in the heterogeneous energy storage unit is controlled, so that the high-power-density buffer unit enters a high-power-ready state. Within the time period corresponding to the high-power pulse load segment data, the real-time power demand data of the LED lighting module is obtained, the real-time power demand data is compared with the preset basic lighting power threshold, and dynamic power allocation instruction data is generated based on the comparison result. In response to the dynamic power allocation command data, the high power density buffer unit is scheduled to output the instantaneous peak power component, and the high energy density main energy storage unit is controlled to output the background power component, wherein the sum of the instantaneous peak power component and the background power component satisfies the real-time demand power data.

[0008] In one embodiment, the step of fusing the intersection traffic phase time-series data and the real-time traffic flow density data to generate predicted lighting load power curve data includes: The traffic phase timing data at the intersection is analyzed, and the data of each phase stage and its corresponding duration for at least one future signal cycle are determined according to the predetermined phase switching schedule of the traffic lights. Based on a pre-trained phase-power mapping model, the corresponding reference lighting power level data is determined for each phase stage data. Based on the pre-trained traffic flow-power correction model, the real-time traffic flow density data is processed to calculate the real-time lighting power correction coefficient data. The reference lighting power level data and the corresponding real-time lighting power correction coefficient data are applied to the preset basic lighting power data to obtain the predicted lighting load power curve data.

[0009] In one embodiment, the step of processing the real-time traffic density data based on a pre-trained traffic flow-power correction model and calculating the real-time lighting power correction coefficient data includes: The real-time traffic density data at the current moment is combined with the historical traffic density data within at least one historical time window to form a traffic density data sequence. The traffic flow density data sequence is input into the pre-trained traffic flow-power correction model so that the pre-trained traffic flow-power correction model outputs the real-time lighting power correction coefficient data.

[0010] In one embodiment, the step of identifying high-power pulse load segment data based on the predicted lighting load power curve data, and generating pre-scheduling instruction data based on the high-power pulse load segment data includes: The predicted lighting load power curve data is processed by moving average filtering to obtain smoothed power curve data; The smoothed power curve data is subjected to point-by-point difference calculation to obtain the power change rate data; In the smoothed power curve data, find continuous data segments where the power value exceeds the first power threshold and the corresponding power change rate data exceeds the first change rate threshold, and use them as high-power pulse candidate segment data; When the duration of the high-power pulse candidate segment data exceeds the shortest pulse duration threshold, the high-power pulse candidate segment data is marked as the high-power pulse load segment data.

[0011] In one embodiment, the method further includes: When at least two high-power pulse load segment data are identified, the interval data between the start time points of two adjacent high-power pulse load segment data is calculated. If the interval data is less than the preset minimum interval threshold, the data of the at least two high-power pulse load segments are merged to generate new high-power pulse load segment data, and the start time data, duration data and peak power data of the new high-power pulse load segment data are updated.

[0012] In one embodiment, the step of controlling the charging and discharging state of the high-power-density buffer unit in the heterogeneous energy storage unit based on the pre-scheduling command data, so that the high-power-density buffer unit enters a high-power-ready state, includes: The pre-scheduling instruction data is parsed to obtain the power buffer start time data and preset advance time data corresponding to the high-power pulse load segment data; At a time when the power buffer start-up time data is reached but is greater than or equal to the preset advance time data, a preparation command is sent to the power conversion circuit connected to the high power density buffer unit. The preparation command includes preparation voltage setting value data, which is used to control the terminal voltage of the high power density buffer unit. In response to the execution of the preparatory command, the output status data of the high power density buffer unit is monitored in real time; Based on the output status data, determine whether the high power density buffer unit is in an effective ready state; If the high power density buffer unit is not in an effective ready state, the ready voltage setting value data is updated according to a preset adjustment strategy, and a new ready command containing the updated ready voltage setting value data is sent so that the high power density buffer unit reaches the high power ready state.

[0013] In one embodiment, the step of determining whether the high power density buffer unit is in an effective ready state based on the output state data includes: Extract steady-state monitoring current data from the output status data; Calculate the deviation rate between the steady-state monitoring current data and the preset target current data; If the deviation rate data exceeds a preset deviation threshold, it is determined that the high power density buffer unit is not in an effective preparation state.

[0014] In one embodiment, the steps of acquiring real-time power demand data of the LED lighting module within the time period corresponding to the high-power pulse load segment data, comparing the real-time power demand data with a preset basic lighting power threshold, and generating dynamic power allocation instruction data based on the comparison result include: Within the time period corresponding to the high-power pulse load segment data, the real-time power demand data of the LED lighting module is obtained at a fixed sampling period; Calculate the absolute value of the difference between the real-time demand power data obtained in the current sampling period and the historical demand power data obtained in the previous sampling period; The standard comparison process is executed, and dynamic power allocation instruction data is generated and output based on the comparison results between the real-time demand power data obtained in the current sampling period and the basic lighting power threshold. If the absolute value of the difference exceeds the preset power mutation threshold, the fast response mode is activated synchronously, and fast response factor data is generated based on the absolute value of the difference. Based on the fast response factor data, the power setting parameters in the dynamic power allocation command data are corrected to obtain the corrected dynamic power allocation command data, which is then overwritten and output.

[0015] In one embodiment, the method further includes: In the fast response mode, the real-time demand power data obtained in subsequent sampling periods is continuously monitored; If the absolute value of the difference data is less than the preset power mutation threshold within a consecutive preset number of sampling periods, then generate mode exit instruction data. In response to the mode exit command data, the system exits the fast response mode and stops modifying the dynamic power allocation command data based on the fast response factor data.

[0016] Furthermore, to achieve the above objectives, this application also proposes a multi-level dynamic scheduling system for photovoltaic energy storage streetlights at urban road intersections. The multi-level dynamic scheduling system for photovoltaic energy storage streetlights at urban road intersections includes: a memory, a processor, and a multi-level dynamic scheduling program for photovoltaic energy storage streetlights at urban road intersections stored in the memory and executable on the processor. The multi-level dynamic scheduling program for photovoltaic energy storage streetlights at urban road intersections is configured to implement the steps of the multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections.

[0017] The proposed method and system for multi-level dynamic scheduling of photovoltaic energy storage streetlights at urban road intersections acquires traffic phase time-series data and real-time traffic flow density data at the intersection, performs fusion processing to generate predicted lighting load power curve data, identifies high-power pulse load segments, controls the charging and discharging state of energy storage units, and dynamically allocates power during high-power periods to meet real-time needs. It can proactively adapt to changes in traffic flow, finely coordinate the operation of energy storage units, effectively extend the lifespan of energy storage units, and improve lighting stability. Attached Figure Description

[0018] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating an embodiment of the multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections in this application. Figure 2 This is a structural schematic diagram of an embodiment of the photovoltaic energy storage street light multi-level dynamic scheduling system at urban road intersections provided in this application.

[0021] Explanation of icon numbers: 10. Memory; 20. Processor.

[0022] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0023] The technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of this application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0024] It should be understood that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0025] In existing technologies, photovoltaic energy storage street light systems face challenges in complex traffic conditions such as urban road intersections. The lighting load exhibits intense, transient, and high-power pulse characteristics, making the average load-based charging and discharging control strategy ineffective in responding to power surges. This leads to accelerated capacity decay and shortened cycle life of the battery under frequent high-current pulse discharges, posing a risk of thermal runaway. Furthermore, insufficient instantaneous output capacity of the battery can cause LED lights to flicker or lose brightness, impacting traffic safety. Existing solutions fail to fundamentally address the damage to energy storage units caused by pulsed loads from an energy dispatching mechanism perspective, thus limiting the system's reliability and economy under high-dynamic load scenarios.

[0026] Based on this, embodiments of this application provide a multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections, referring to... Figure 1 The multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections includes steps S100 to S500, wherein: Step S100: Obtain traffic phase time series data and real-time traffic flow density data at the intersection; perform fusion processing on the traffic phase time series data and the real-time traffic flow density data at the intersection to generate predicted lighting load power curve data. Step S200: Based on the predicted lighting load power curve data, identify high-power pulse load segment data, and generate pre-schedule instruction data according to the high-power pulse load segment data; Step S300: Based on the pre-scheduling instruction data, control the charging and discharging state of the high power density buffer unit in the heterogeneous energy storage unit, so that the high power density buffer unit enters the high power ready state. Step S400: Within the time period corresponding to the high-power pulse load segment data, obtain the real-time power demand data of the LED lighting module, compare the real-time power demand data with the preset basic lighting power threshold, and generate dynamic power allocation instruction data based on the comparison result. Step S500: In response to the dynamic power allocation command data, the high power density buffer unit is scheduled to output the instantaneous peak power component, and the high energy density main energy storage unit is controlled to output the background power component, wherein the sum of the instantaneous peak power component and the background power component satisfies the real-time demand power data.

[0027] In this embodiment, a heterogeneous energy storage unit refers to an energy storage system composed of two or more different types of energy storage devices. In this embodiment, it typically includes a high-power-density buffer unit and a high-energy-density main energy storage unit, aiming to combine the advantages of different energy storage devices to address the complex characteristics of lighting loads at urban road intersections. The high-power-density buffer unit refers to an energy storage device with rapid charging and discharging capabilities and high instantaneous power output capabilities, such as a supercapacitor or a high-rate lithium-ion battery. This unit is mainly used to absorb and release short-term, high-power pulse loads to protect the high-energy-density main energy storage unit and improve system response speed. The high-energy-density main energy storage unit refers to an energy storage device with high energy storage density and long-term continuous discharge capabilities, such as a conventional lithium-ion battery pack or lead-acid battery pack. This unit is mainly responsible for providing the system's basic lighting power and long-term energy supply, and serves as a supplementary energy source for the high-power-density buffer unit.

[0028] In this embodiment, the high-power ready state refers to a preparatory working state in which the high-power density buffer unit, after receiving a pre-scheduling instruction, adjusts its internal operating parameters to quickly respond to the upcoming high-power pulse load and output instantaneous peak power with maximum efficiency. The high-power pulse load segment data refers to the load interval in the predicted lighting load power curve data where the power value increases rapidly and persists for a period of time. This data segment reflects the characteristic of a sharp increase in lighting power demand at urban road intersections during traffic signal switching or when traffic flow is dense. The predicted lighting load power curve data refers to the expected power demand trend of LED lighting modules over a future period, obtained by fusing traffic phase time-series data and real-time traffic density data at intersections. This curve data provides a basis for the system to perform forward-looking scheduling.

[0029] In this embodiment, the multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections first acquires traffic phase time-series data and real-time traffic flow density data at the intersection. The traffic phase time-series data can be periodically released by traffic management departments or obtained through manual input. Real-time traffic flow density data can be collected in real time by sensors installed at the intersection (e.g., geomagnetic sensors, radar sensors, or video analysis systems). This data is transmitted to the processing unit. Subsequently, the traffic phase time-series data and the real-time traffic flow density data are fused to generate a predicted lighting load power curve. The fusion process can be simply achieved by linearly superimposing or weighting the traffic phase time-series data with the real-time traffic flow density data to roughly estimate the future lighting load. For example, the average lighting power demand for a future time period can be directly calculated based on the traffic light duration and the current average traffic flow, and this can be used to generate the predicted lighting load power curve.

[0030] Furthermore, based on the predicted lighting load power curve data, high-power pulse load segments are identified, and pre-schedule instruction data is generated according to these high-power pulse load segment data. The predicted lighting load power curve data is input to the analysis module. High-power pulse load segments can be identified by setting a fixed power threshold; that is, when the predicted power value exceeds the threshold, it is determined to be a high-power pulse load segment. For example, an empirical power upper limit can be set, and any continuous power value exceeding the upper limit is marked as a pulse segment. Once the high-power pulse load segment data is identified, the system generates pre-schedule instruction data containing parameters such as its start time, duration, and peak power. This instruction data is then sent to the energy storage management system.

[0031] Based on this, and using the pre-scheduling command data, the charging and discharging state of the high-power-density buffer unit in the heterogeneous energy storage unit is controlled, enabling the high-power-density buffer unit to enter a high-power-ready state. The pre-scheduling command data is received by the energy storage management system. According to the time indicated in the command when the pulse load is about to arrive, the system sends a simple start signal to the high-power-density buffer unit in advance. This start signal prompts the power conversion circuit of the high-power-density buffer unit to enter a working state, for example, by performing a small amount of charging or discharging through a preset fixed voltage or current value, bringing its internal state to a basic preparation level, thereby enabling it to respond to subsequent power demands.

[0032] In this embodiment, during the time period corresponding to the high-power pulse load segment, real-time power demand data of the LED lighting module is acquired. This real-time power demand data is compared with a preset basic lighting power threshold, and dynamic power allocation command data is generated based on the comparison result. During the time period corresponding to the high-power pulse load segment, the real-time power demand data of the LED lighting module is continuously measured using current and voltage sensors. This real-time power demand data is transmitted to the controller. The controller directly compares the real-time power demand data with a preset basic lighting power threshold. For example, if the real-time power demand data is higher than the threshold, a command is generated indicating that additional power output is needed; if it is lower than or equal to the threshold, a command is generated to maintain the basic power output. Based on this simple comparison result, dynamic power allocation command data is generated, which includes the power allocation requirements for the energy storage unit.

[0033] Therefore, in response to the dynamic power allocation command data, the high-power-density buffer unit is scheduled to output the instantaneous peak power component, and the high-energy-density main energy storage unit is controlled to output the background power component. The sum of the instantaneous peak power component and the background power component satisfies the real-time power demand data. The dynamic power allocation command data is received by the energy storage management system. According to the command, the system controls the power converter of the high-power-density buffer unit to quickly respond and output the portion of the real-time power demand data exceeding the basic lighting power threshold, i.e., the instantaneous peak power component. Simultaneously, the power converter of the high-energy-density main energy storage unit is controlled to provide the background power component represented by the basic lighting power threshold. Through the coordinated work of the high-power-density buffer unit and the high-energy-density main energy storage unit, the sum of their output instantaneous peak power component and background power component can accurately meet the real-time power demand data of the LED lighting module. This collaborative scheduling method allows the high-power-density buffer unit to focus on handling instantaneous high-power surges, while the high-energy-density main energy storage unit is responsible for stable, long-term energy supply.

[0034] In this embodiment, by proactively predicting the lighting load pulses at urban road intersections and pre-emptively putting the high-power-density buffer unit into a high-power-ready state, a rapid response to instantaneous high-power demands is achieved. Consequently, the high-power-density buffer unit can effectively handle instantaneous peak power, reducing the discharge pressure on the high-energy-density main energy storage unit, thereby extending its cycle life and reducing the risk of thermal runaway. Simultaneously, this method ensures that the LED lighting module still receives a stable and sufficient power supply when traffic conditions change drastically, avoiding flickering or brightness reduction, and improving traffic safety and system reliability.

[0035] In one feasible implementation, the step of fusing the intersection traffic phase time-series data and the real-time traffic flow density data to generate predicted lighting load power curve data includes: parsing the intersection traffic phase time-series data; determining the phase stage data and its corresponding duration data for at least one future signal cycle according to the predetermined phase switching schedule of traffic lights; determining the corresponding reference lighting power level data for each phase stage data based on a pre-trained phase-power mapping model; processing the real-time traffic flow density data based on a pre-trained traffic flow-power correction model to calculate the real-time lighting power correction coefficient data; and applying the reference lighting power level data and the corresponding real-time lighting power correction coefficient data to preset basic lighting power data to obtain the predicted lighting load power curve data.

[0036] In this embodiment, the traffic phase timing data of the intersection is parsed, and according to the predetermined phase switching schedule of the traffic lights, the data of each phase stage and its corresponding duration for at least one future signal cycle are determined. This step aims to acquire and understand the structured information of the traffic lights operating at the intersection. Intersection traffic phase timing data typically includes the cycle length of the traffic lights, the order of each phase, the duration of each phase, and the logic of phase switching. By parsing this data, the system can anticipate how traffic flow will be organized in one or more future signal cycles, such as which directions of vehicles are allowed to pass and which directions need to wait. This provides a basic time frame for predicting lighting requirements under different traffic conditions. For example, this data can be obtained directly from the traffic management department's signal control system or acquired in real time through communication with the traffic signal controller.

[0037] Based on this, a pre-trained phase-power mapping model is used to determine the corresponding baseline lighting power level for each phase phase. The phase-power mapping model is a pre-trained model using historical data, and its core function is to output a corresponding baseline lighting power level based on different traffic phases. By learning from historical data, this model can identify the lighting intensity typically required at intersections under specific traffic phases. For example, a higher lighting power level may be needed on main roads with high traffic volume to ensure driving safety and comfort; while a lower lighting power level may be required on secondary roads with lower traffic volume or pedestrian-only phases. This model can be trained using various machine learning algorithms, such as regression analysis or neural networks, on historical traffic phase data and actual lighting power data.

[0038] Simultaneously, based on a pre-trained traffic flow-power correction model, the real-time traffic density data is processed to calculate real-time lighting power correction coefficients. This traffic flow-power correction model is also a pre-trained model; its function is to calculate a lighting power correction coefficient based on real-time traffic density data. Real-time traffic density data reflects the current density of vehicles at the intersection. This correction model can capture the dynamic relationship between traffic density and lighting demand. For example, when traffic density is high, higher lighting brightness may be needed to cope with more complex traffic conditions and drivers' visual needs; conversely, when traffic density is low, lighting brightness can be appropriately reduced to save energy. This model can be obtained by training on historical traffic density data and actual lighting power correction demand data, for example, using time series analysis or deep learning methods.

[0039] In this embodiment, the reference lighting power level data and the corresponding real-time lighting power correction coefficient data are applied to the preset base lighting power data to obtain the predicted lighting load power curve data. This step integrates the aforementioned determined reference lighting power level data and real-time lighting power correction coefficient data to generate the final predicted lighting load power curve data. The preset base lighting power data can be understood as the default lighting power without any correction or specific phase considerations. By applying the reference lighting power level data (reflecting phase influence) and the real-time lighting power correction coefficient data (reflecting traffic density influence) to the base lighting power data, a refined and dynamically changing predicted lighting load power curve data can be obtained. For example, the base lighting power data can be multiplied by the reference lighting power level data (which may be a scaling factor) and then multiplied by the real-time lighting power correction coefficient data, or a more complex functional relationship can be used for calculation to obtain the expected power demand trend of the street lighting system in the future time period.

[0040] In this embodiment, by analyzing traffic phase time-series data, the system can predict future traffic flow patterns, thus providing a structured time frame and baseline power level for lighting needs under different traffic conditions. Simultaneously, real-time traffic density data is introduced and corrected through a pre-trained model, enabling the prediction results to dynamically adapt to changes in actual traffic flow at the current intersection. This prediction method, combining the structure of traffic phases and the dynamics of real-time traffic flow, makes the generated predicted lighting load power curve data more accurate and closer to actual needs. Accurate predicted load data provides a reliable basis for subsequent scheduling of heterogeneous energy storage units, avoiding energy waste or insufficient power supply due to inaccurate predictions. This improves the operating efficiency and reliability of the entire photovoltaic energy storage street light system, ensuring appropriate lighting under different traffic scenarios while optimizing energy management.

[0041] In one feasible implementation, the step of processing the real-time traffic density data based on a pre-trained traffic flow-power correction model to calculate the real-time lighting power correction coefficient data includes: combining the real-time traffic density data at the current moment with historical traffic density data within at least one historical time window to form a traffic density data sequence; and inputting the traffic density data sequence into the pre-trained traffic flow-power correction model so that the pre-trained traffic flow-power correction model outputs the real-time lighting power correction coefficient data.

[0042] In this embodiment, real-time traffic density data at the current moment is combined with historical traffic density data within at least one historical time window to form a traffic density data sequence, aiming to provide richer temporal dimension information for subsequent model processing. Real-time traffic density data refers to the density of vehicles on the road detected by sensors (e.g., geomagnetic coils, video analysis systems, radar sensors, etc.) at the current moment. Historical traffic density data refers to vehicle density data recorded before the current moment within a preset time window (e.g., the past 5 minutes, 10 minutes, or a longer period). By arranging and combining these data in chronological order, a sequence with temporal context is formed, for example, a vector containing traffic density values ​​at the current moment and the previous N sampling moments. This serialized data can reflect the dynamic trend of traffic flow changes, rather than just the instantaneous state.

[0043] In this embodiment, the traffic flow density data sequence is input into the pre-trained traffic flow-power correction model, so that the pre-trained traffic flow-power correction model outputs the real-time lighting power correction coefficient data. The traffic flow-power correction model is an offline-trained machine learning model, whose training data typically includes a large number of historical traffic flow density data sequences and their corresponding actual lighting power requirements or correction coefficients. This model can employ various sequence processing algorithms, such as recurrent neural networks (RNNs), long short-term memory networks (LSTMs), gated recurrent units (GRUs), or attention-based Transformer models. When a traffic flow density data sequence is received as input, the model analyzes and processes the sequence data based on the complex patterns and mapping relationships learned during the training phase, thereby predicting and outputting a real-time lighting power correction coefficient. This correction coefficient is typically a floating-point number between 0 and 1, used to adjust the base lighting power to adapt to actual traffic conditions.

[0044] In this embodiment, through the above technical solution, this application can provide input data with more temporal continuity and contextual information for the pre-trained traffic flow-power correction model. This enables the model to not only perceive the current traffic conditions but also understand the evolution trend and historical patterns of traffic flow, thereby improving the prediction accuracy of the lighting power correction coefficient and the robustness of the model. Compared with solutions that rely solely on instantaneous data, this solution can adjust lighting power more precisely and in a more timely manner, avoiding excessive or insufficient lighting caused by instantaneous fluctuations in traffic flow. This optimizes energy utilization efficiency and enhances the adaptability of road lighting and the user experience.

[0045] In one feasible implementation, the step of identifying high-power pulse load segment data based on the predicted lighting load power curve data and generating pre-scheduling instruction data based on the high-power pulse load segment data includes: performing moving average filtering on the predicted lighting load power curve data to obtain smoothed power curve data; performing point-by-point difference calculation on the smoothed power curve data to obtain power change rate data; in the smoothed power curve data, finding continuous data segments where the power value exceeds a first power threshold and the corresponding power change rate data exceeds the first change rate threshold, and using these segments as high-power pulse candidate segment data; when the duration of the high-power pulse candidate segment data exceeds the shortest pulse duration threshold, marking the high-power pulse candidate segment data as the high-power pulse load segment data.

[0046] In this embodiment, the predicted lighting load power curve data is first processed by moving average filtering to eliminate random noise and short-term fluctuations in the data, thereby obtaining smoother power curve data that better reflects the true trend. Moving average filtering can be implemented by setting a sliding window of fixed length; for example, the average of the current point and its N previous points can be calculated as the filtering result for the current point. This processing helps reduce misjudgments in subsequent identification processes and improves data reliability. Then, point-by-point difference calculation is performed on the smoothed power curve data to obtain the power change rate data. The power change rate data can intuitively reflect the speed and direction of power change over time, which is crucial for identifying rapid increases or decreases in power. For example, it can be obtained by calculating the ratio of the power difference between two adjacent sampling points to the time interval, or by using other numerical differentiation methods.

[0047] Based on this, the system searches for continuous data segments that meet specific conditions within the smoothed power curve data, designating them as candidate segments for high-power pulses. These conditions include: the power value exceeding a preset first power threshold, and the corresponding power change rate exceeding a preset first change rate threshold. The first power threshold can be set according to the actual application scenario and the rated power of the streetlights; for example, it can be set as a percentage of the basic lighting power or a fixed value. The first change rate threshold is used to ensure that the identified power change is sufficient, rather than a minor fluctuation. By considering both the absolute value and the rate of change of power simultaneously, the true characteristics of power pulses can be captured more accurately, avoiding misjudgments that may result from relying solely on a single indicator such as power magnitude or rate of change.

[0048] In this embodiment, to further ensure that the identified high-power pulses are valid and require pre-scheduling, the system determines the duration of the high-power pulse candidate segment data. Only when its duration exceeds a preset minimum pulse duration threshold will the candidate segment be ultimately marked as high-power pulse load segment data. The minimum pulse duration threshold can be determined based on actual traffic flow characteristics and the response time of the energy storage unit, for example, set to several seconds to tens of seconds. This step effectively avoids the situation where instantaneous spikes or brief fluctuations are incorrectly identified as high-power pulses, ensuring the robustness of the identification results, so that only power demands with a sufficiently long duration are considered high-power pulses.

[0049] In this embodiment, through the above technical solution, this application can effectively solve the noise and fluctuation problems existing in the predicted lighting load power curve data, avoiding misjudgment or omission of high-power pulse loads due to inaccurate data. By using moving average filtering, data noise is effectively suppressed, allowing the true trend of the power curve to be revealed. Combining a dual judgment mechanism of power value and power change rate, the rising and maintaining states of power can be accurately captured. Furthermore, by filtering through a duration threshold, short-term, non-substantial power spikes are eliminated, ensuring that the identified high-power pulse load segment data is real, stable, and requires pre-scheduling. This makes the generated pre-scheduling command data more accurate and reliable, thereby enabling more effective control of the charging and discharging state of the high-power density buffer units in the heterogeneous energy storage units, allowing them to accurately enter a high-power ready state before the actual high-power pulse load arrives. Ultimately, this improves the response efficiency, stability, and energy utilization efficiency of the multi-level dynamic scheduling system for photovoltaic energy storage streetlights at urban road intersections.

[0050] In one feasible implementation, the method further includes: when at least two high-power pulse load segment data are identified, calculating the interval data between the start time points of two adjacent high-power pulse load segment data; if the interval data is less than a preset minimum interval threshold, merging the at least two high-power pulse load segment data to generate new high-power pulse load segment data, and updating the start time data, duration data, and peak power data of the new high-power pulse load segment data.

[0051] In this embodiment, after the aforementioned filtering, differential analysis, threshold judgment, and duration screening steps, the system has initially identified a series of independent high-power pulse load segments. These pulse segments represent scenarios where the LED lighting module may require instantaneous high-power output within a specific time window. The phrase "at least two" emphasizes the prerequisite for subsequent merging operations, namely, the existence of multiple potential merging targets. When at least two high-power pulse load segment data are identified, the system calculates the interval data between the start times of two adjacent high-power pulse load segment data. The interval data refers to the duration between the start times of two consecutively identified high-power pulse load segments on the time axis. This interval data can be obtained through simple time difference calculation, for example, by subtracting the start time of the previous pulse segment from the start time of the later pulse segment. The purpose of this step is to quantify the closeness of adjacent pulse segments, providing key parameters for subsequent merging decisions.

[0052] In this embodiment, the calculated interval data is then compared with a preset minimum interval threshold. The minimum interval threshold is a pre-set time value used to determine whether two adjacent high-power pulse load segments are close enough to require merging. This threshold can be empirically set or determined through simulation optimization based on factors such as the actual system's response speed, the granularity of the scheduling strategy, and the tolerance for scheduling overhead. For example, if the system is inefficient in frequently switching scheduling instructions within a short period, a relatively large minimum interval threshold can be set. If the interval data is less than the preset minimum interval threshold, the data of at least two high-power pulse load segments are merged to generate new high-power pulse load segment data. The merging operation refers to treating multiple time-closely connected high-power pulse load segments as a single, longer pulse event. The specific merging method can be to logically connect the time ranges of these pulse segments to form a continuous time interval. After merging, the system updates the start time data, duration data, and peak power data of the new high-power pulse load segment data. After merging, it is necessary to recalculate and update the key parameters of the newly generated high-power pulse load segment. The start time data is typically taken as the earliest start time among all pulse segments before merging; the duration data is the difference between the end time and start time of the new pulse segment after merging; the peak power data can be the maximum peak power among all pulse segments before merging, or the maximum power value can be recalculated over the entire time interval after merging. These updates ensure that the merged pulse segment data accurately reflects the actual load demand and provides accurate input for subsequent pre-scheduling instructions.

[0053] In this embodiment, through the above technical solution, when the system identifies multiple high-power pulse load segments that are closely connected in time, it can intelligently merge them into a longer and more comprehensive high-power pulse load segment. This merging mechanism effectively avoids the system overhead and scheduling fragmentation problems caused by frequent scheduling within short time intervals. By calculating the interval data of adjacent pulse segments and comparing it with a preset minimum interval threshold, it ensures that merging only occurs when truly needed, thereby improving the robustness and efficiency of the scheduling strategy. The merged new high-power pulse load segment data has updated start time, duration, and peak power, which can more accurately reflect the actual continuous high-power demand. This allows heterogeneous energy storage units, especially high-power density buffer units, to prepare and schedule for a longer and more concentrated high-power event, thereby ensuring a stable and sufficient instantaneous peak power component during the entire high-power demand period. This avoids power instability or response delays caused by frequent switching or insufficient preparation, improving the reliability and energy efficiency of urban road intersection lighting systems.

[0054] In one feasible implementation, the step of controlling the charging and discharging state of a high-power-density buffer unit in a heterogeneous energy storage unit based on the pre-scheduling instruction data, so that the high-power-density buffer unit enters a high-power-ready state, includes: parsing the pre-scheduling instruction data to obtain power buffer start-up time data and preset advance time data corresponding to the high-power pulse load segment data; sending a preparatory instruction to the power conversion circuit connected to the high-power-density buffer unit at a time before reaching the power buffer start-up time data and greater than or equal to the preset advance time data, the preparatory instruction including preparatory voltage setting value data for controlling the terminal voltage of the high-power-density buffer unit; monitoring the output status data of the high-power-density buffer unit in real time in response to the execution of the preparatory instruction; determining whether the high-power-density buffer unit is in an effective preparatory state based on the output status data; if the high-power-density buffer unit is not in an effective preparatory state, updating the preparatory voltage setting value data according to a preset adjustment strategy, and sending a new preparatory instruction including the updated preparatory voltage setting value data, so that the high-power-density buffer unit reaches the high-power-ready state.

[0055] In this embodiment, the pre-scheduled command data is a control command generated in advance by the system based on the predicted lighting load power curve data, especially the identified high-power pulse load segment data. The purpose of parsing this command is to extract key timing information, including "power buffer start-up time data" indicating when the high-power density buffer unit needs to begin preparation, and "preset advance time data" indicating the time margin required to ensure adequate preparation. This timing data forms the basis for subsequent precise control of the high-power density buffer unit to enter the ready state, ensuring it has sufficient time to adjust its state before the actual high-power demand arrives.

[0056] In this embodiment, before the predetermined power buffer start-up time data arrives, the system calculates a suitable timing based on preset advance time data and sends a preparatory command to the power conversion circuit of the high power density buffer unit. This power conversion circuit is typically a DC / DC converter or inverter, responsible for managing the charging and discharging process of the high power density buffer unit (e.g., a supercapacitor or a high-rate lithium battery). The preparatory voltage setting value included in the preparatory command is predetermined based on the characteristics of the high power density buffer unit and the expected high power output requirements. It is used to adjust the terminal voltage of the high power density buffer unit to a specific level that facilitates rapid response and efficient output of instantaneous peak power. By precisely controlling the terminal voltage, the internal resistance and response speed of the high power density buffer unit can be optimized.

[0057] In this embodiment, once the preparatory command is sent and execution begins, the system continuously and in real-time monitors the output status data of the high-power-density buffer unit. This output status data may include, but is not limited to, parameters such as the high-power-density buffer unit's terminal voltage, charging / discharging current, internal temperature, and state of charge (SOC). The purpose of real-time monitoring is to obtain the current operating status of the high-power-density buffer unit in order to assess whether it has entered a valid preparatory state as required by the preparatory command, and to provide a basis for subsequent judgments and adjustments.

[0058] In this embodiment, the system compares the real-time monitored output status data with preset effective ready-to-go standards to determine whether the high-power-density buffer unit has truly reached a "high-power ready state." The standards for an effective ready-to-go state may include whether the terminal voltage is within the target range, whether the charging and discharging current is stable, and whether the internal temperature is suitable. This determination step is crucial to ensuring that the high-power-density buffer unit can reliably provide instantaneous peak power, avoiding scheduling failures due to poor status.

[0059] In this embodiment, if the judgment result indicates that the high-power-density buffer unit is not in an effective ready state, the system will initiate a feedback adjustment mechanism. Based on a preset adjustment strategy (e.g., PID control, fuzzy control, or lookup table method), the system will calculate and update the ready voltage setpoint data. The updated ready voltage setpoint data will be sent to the power conversion circuit via a new ready command to further fine-tune the terminal voltage of the high-power-density buffer unit until it reaches or approaches the ideal high-power ready state. This closed-loop control ensures that the high-power-density buffer unit can always maintain optimal responsiveness and output performance when facing actual high-power pulse loads.

[0060] In this embodiment, through the above technical solution, this application can effectively solve the problem that high power density buffer units may be unable to respond in a timely and stable manner due to poor condition when facing instantaneous high power pulse loads. By parsing the pre-scheduling command data and obtaining the power buffer start-up time data and preset advance time data in advance, the system can send a preparatory command containing the preparatory voltage setpoint data to the power conversion circuit of the high power density buffer unit before the critical moment arrives, thereby pre-adjusting its terminal voltage. On this basis, the system monitors the output status data of the high power density buffer unit in real time and determines whether it is in an effective preparatory state based on the data. If it is found that it is not in an effective preparatory state, the system can update the preparatory voltage setpoint data in a timely manner and send a new preparatory command for correction according to the preset adjustment strategy. This combination of forward-looking preparatory control and real-time feedback adjustment forms a closed-loop, adaptive preparation mechanism, ensuring that the high power density buffer unit can reliably and efficiently output the instantaneous peak power component within the time period corresponding to the actual high power pulse load segment data. This improves the response speed and stability of photovoltaic energy storage street light systems at urban road intersections to fluctuations in lighting load power when traffic flow changes drastically, ensuring lighting quality and extending the service life of energy storage units.

[0061] In one feasible implementation, the step of determining whether the high power density buffer unit is in an effective preparation state based on the output state data includes: extracting steady-state monitoring current data from the output state data; calculating the deviation rate data between the steady-state monitoring current data and the preset target preparation current data; and determining that the high power density buffer unit is not in an effective preparation state if the deviation rate data exceeds a preset deviation threshold.

[0062] In this embodiment, extracting steady-state monitoring current data from the output status data refers to the process where, after the preparatory command is sent, the high-power-density buffer unit begins to adjust its internal state to achieve a high-power-ready state. During this process, the system monitors various electrical parameters of the buffer unit in real time, generating output status data. Steady-state monitoring current data specifically refers to the current value that represents the current stable output capability of the buffer unit, obtained by filtering or calculating from these real-time monitoring data after the preparatory command is executed and a short stabilization period has elapsed. For example, after the preparatory command is sent, a preset stabilization delay time (e.g., several milliseconds to tens of milliseconds) can be waited, and then the current data within that time period can be continuously sampled. The sampled values ​​can then be averaged or low-pass filtered to eliminate transient fluctuations and noise interference, thereby obtaining a more accurate steady-state current value. This step aims to obtain the true and stable operating current information of the high-power-density buffer unit in the preparatory state, providing reliable input for subsequent performance evaluation.

[0063] In this embodiment, calculating the deviation rate between the steady-state monitoring current data and the preset target preparation current data involves comparing the extracted steady-state monitoring current data with the current output value that the high-power-density buffer unit should achieve in its effective preparation state (i.e., the preset target preparation current data). The preset target preparation current data is an ideal current value determined during the system design or commissioning phase based on factors such as the rated parameters of the high-power-density buffer unit, the system's requirements for instantaneous peak power response, and actual operating experience. The deviation rate can be obtained by calculating the relative difference between the steady-state monitoring current data and the preset target preparation current data, for example, using the formula "(steady-state monitoring current data - preset target preparation current data) / preset target preparation current data × 100%". This step provides a quantitative indicator to measure the degree of deviation between the actual preparation state of the high-power-density buffer unit and the expected target.

[0064] In this embodiment, if the deviation rate exceeds a preset deviation threshold, the high-power-density buffer unit is determined not to be in an effective ready state. The preset deviation threshold is an acceptable deviation range that defines the maximum permissible deviation between the actual operating current and the target current of the high-power-density buffer unit. This threshold can be set according to the specific requirements of the system regarding power output accuracy, response speed, and stability; for example, it can be set to ±5% or ±10%. When the calculated deviation rate exceeds this preset threshold, it indicates that the current ready state of the high-power-density buffer unit fails to meet the system's requirements for a high-power-ready state, and is therefore determined not to be in an effective ready state. This determination will serve as the basis for subsequent adjustment strategies.

[0065] In this embodiment, by extracting steady-state monitoring current data from the output status data and calculating the deviation rate between it and the preset target preparation current data, the matching degree between the actual working state and the expected state of the buffer unit can be quantitatively evaluated. When the deviation rate data exceeds the preset deviation threshold, the system can promptly and accurately determine that the buffer unit is not in an effective preparation state, thereby triggering subsequent adjustment strategies. This avoids misjudgments that may be caused by relying solely on voltage settings or simple monitoring, ensuring that the high-power-density buffer unit can quickly and accurately provide instantaneous peak power components when needed, improving the dynamic response capability and power supply reliability of the entire photovoltaic energy storage street light system. Especially when dealing with complex and variable traffic load pulses at urban road intersections, it can more effectively maintain lighting quality and system stability.

[0066] In one feasible implementation, the steps of acquiring real-time power demand data of the LED lighting module within the time period corresponding to the high-power pulse load segment data, comparing the real-time power demand data with a preset basic lighting power threshold, and generating dynamic power allocation instruction data based on the comparison result include: acquiring real-time power demand data of the LED lighting module at a fixed sampling period within the time period corresponding to the high-power pulse load segment data; calculating the absolute value of the difference between the real-time power demand data obtained in the current sampling period and the historical power demand data obtained in the previous sampling period; executing a conventional comparison process, generating and outputting dynamic power allocation instruction data based on the comparison result between the real-time power demand data obtained in the current sampling period and the basic lighting power threshold; if the absolute value of the difference exceeds a preset power mutation threshold, simultaneously activating a fast response mode, generating fast response factor data based on the absolute value of the difference; and correcting the power setting parameters in the dynamic power allocation instruction data based on the fast response factor data to obtain corrected dynamic power allocation instruction data and overwriting the output.

[0067] In this embodiment, within the time period corresponding to the high-power pulse load segment data, the system continuously acquires the real-time power demand data of the LED lighting module at a preset fixed sampling period. This fixed sampling period can be set according to the system's response speed requirements; for example, it can be set to 100 milliseconds, 200 milliseconds, or 500 milliseconds to ensure continuous and detailed monitoring of the actual power demand of the lighting module. Real-time power demand data can be obtained by directly measuring the actual power consumption of the LED lighting module or through the output power command of its controller. Then, the absolute value of the difference between the real-time power demand data obtained in the current sampling period and the historical power demand data obtained in the previous sampling period is calculated. This step aims to quantify the magnitude of change in the LED lighting module's power demand within adjacent sampling periods. By calculating the absolute value of the difference, the degree of drastic power change can be uniformly measured regardless of whether the power increases or decreases. For example, if the power demand in the current period is 100W and in the previous period it was 80W, the absolute value of the difference is 20W; if the power demand in the previous period was 120W, the absolute value of the difference is also 20W.

[0068] In this embodiment, under normal operation, the system executes a conventional comparison process. Based on the comparison result between the real-time demand power data obtained in the current sampling period and the basic lighting power threshold, dynamic power allocation instruction data is generated and output. This conventional comparison process typically involves comparing the real-time demand power with a preset basic lighting power threshold. For example, when the real-time demand power is higher than the threshold, the energy storage unit is instructed to provide more power; when it is lower than the threshold, the power output is reduced. The generated dynamic power allocation instruction data includes specific power setting parameters to guide how the high-power-density buffer unit and the high-energy-density main energy storage unit in the heterogeneous energy storage unit work together to meet the current lighting demand.

[0069] In this embodiment, to cope with emergencies, if the absolute value of the difference calculated above exceeds a preset power surge threshold, the system will simultaneously activate a fast response mode. The power surge threshold is a pre-set value used to define what level of power change is considered a "surge." For example, it can be set to a power change exceeding 10% or 20W of the total power within a short period. Once a power surge is detected, the system generates fast response factor data based on this absolute value of the difference. The fast response factor data can be a proportionality coefficient, an additional power value, or a time constant, and its magnitude is positively correlated with the amplitude of the power surge, indicating the need for a faster and larger adjustment of the power output.

[0070] In this embodiment, the power setting parameters in the dynamic power allocation command data are finally corrected based on the fast response factor data to obtain the corrected dynamic power allocation command data, which is then overridden and output. This means that in fast response mode, the original power setting parameters will be adjusted by the fast response factor. For example, if the fast response factor is a proportional coefficient, it may increase or decrease the target power output by a larger percentage in a short period of time; if it is an additional power value, it will directly increase or decrease an additional amount of power on the original basis. By overriding the output, it is ensured that the corrected command can take effect immediately, thereby enabling the energy storage system to respond to sudden changes in lighting demand at a faster speed and with a greater magnitude.

[0071] In this embodiment, by introducing fixed-period sampling of real-time power demand data and calculation of the absolute value of the difference, this application can effectively monitor the instantaneous changes in the power demand of LED lighting modules. When a power surge is detected, the system can simultaneously activate a fast response mode and generate a fast response factor based on the magnitude of the power surge, thereby correcting the power setting parameters in the dynamic power allocation command. This mechanism enables the system to respond quickly and accurately to fluctuations in lighting demand caused by sudden changes in traffic flow or ambient light at urban road intersections, avoiding poor lighting effects or energy waste caused by the lag in conventional comparison processes, ensuring the stability and reliability of the lighting system, and optimizing energy utilization efficiency.

[0072] In one feasible implementation, the method further includes: continuously monitoring the real-time demand power data obtained in subsequent sampling periods under the fast response mode; if the absolute value of the difference data is less than a preset power mutation threshold within a consecutive preset number of sampling periods, generating mode exit instruction data; and exiting the fast response mode in response to the mode exit instruction data, and ceasing the correction of the dynamic power allocation instruction data based on the fast response factor data.

[0073] In this embodiment, after the fast response mode is activated, the system continuously acquires real-time power demand data of the LED lighting module at the same frequency as the normal sampling period. This is to ensure that the system can keep abreast of the latest load changes, providing a continuous and accurate data basis for subsequent decisions on whether to exit the fast response mode. This continuously monitored real-time power demand data is received and stored by the system's internal controller or processing unit for further analysis and comparison.

[0074] Based on this, if the absolute value of the difference data is less than a preset power fluctuation threshold within a consecutive preset number of sampling periods, a mode exit command is generated. To implement this judgment logic, the system maintains an internal counter. In each sampling period, the system calculates the absolute value of the difference between the currently obtained real-time demand power data and the historical demand power data obtained in the previous sampling period. If the absolute value of the difference is less than the preset power fluctuation threshold, it indicates that the current power fluctuation is within an acceptable stable range, and the counter increments. Conversely, if the absolute value of the difference is greater than or equal to the power fluctuation threshold, it indicates that power fluctuation still exists, and the counter is reset to zero. When the value of the counter reaches a preset number, it means that the system has been in a relatively stable state for several consecutive sampling periods, and the system generates a mode exit command. This preset number can be configured according to the actual application scenario and the requirements for system stability, for example, set to 3, 5 or more sampling periods, to ensure that the system exits the fast response mode only after it is truly stable, avoiding frequent mode switching due to brief false stability.

[0075] In this embodiment, in response to the mode exit command data, the system will exit the fast response mode and stop modifying the dynamic power allocation command data based on the fast response factor data. When the system receives the mode exit command data, its internal fast response mode status flag will be reset, thereby switching the system back from fast response mode to the normal power allocation mode. Simultaneously, the modification logic previously performed on the dynamic power allocation command data based on the fast response factor data in fast response mode will be disabled or stopped. This means that subsequent dynamic power allocation command data will no longer be affected by the fast response factor, but will be generated and output entirely based on the conventional comparison process (i.e., the comparison result between real-time demand power data and the basic lighting power threshold).

[0076] In this embodiment, through the above technical solution, this application can accurately identify when instantaneous power fluctuations have stabilized and promptly generate mode exit command data, thereby responsively exiting the fast response mode and ceasing the correction of dynamic power allocation command data based on fast response factor data. This effectively avoids the system unnecessarily maintaining a fast response state after the power fluctuation ends, improves the stability of the scheduling strategy and resource utilization efficiency, reduces unnecessary system overhead and scheduling oscillations, and ensures a smooth transition in power allocation and the overall reliability of the system.

[0077] In the embodiments of this application, the multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections acquires traffic phase time-series data and real-time traffic flow density data at the intersection, performs fusion processing to generate predicted lighting load power curve data, identifies high-power pulse load segments, controls the charging and discharging state of energy storage units, and dynamically allocates power during high-power periods to meet real-time needs. It can proactively adapt to changes in traffic flow, finely coordinate the operation of energy storage units, effectively extend the lifespan of energy storage units, and improve lighting stability.

[0078] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections in this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0079] This application also provides a multi-level dynamic scheduling system for photovoltaic energy storage streetlights at urban road intersections, referenced. Figure 2 The photovoltaic energy storage street light multi-level dynamic scheduling system at urban road intersections includes: a memory 10, a processor 20, and a photovoltaic energy storage street light multi-level dynamic scheduling program for urban road intersections stored on the memory 10 and executable on the processor 20. The photovoltaic energy storage street light multi-level dynamic scheduling program for urban road intersections is configured to implement the steps of the photovoltaic energy storage street light multi-level dynamic scheduling method for urban road intersections.

[0080] The multi-level dynamic scheduling system for photovoltaic energy storage streetlights at urban road intersections provided in this application adopts the multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections in the above embodiments, which can extend the lifespan of energy storage units and improve lighting stability. Compared with the prior art, the beneficial effects of the multi-level dynamic scheduling system for photovoltaic energy storage streetlights at urban road intersections provided in this application are the same as the beneficial effects of the multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections provided in the above embodiments, and other technical features of the multi-level dynamic scheduling system for photovoltaic energy storage streetlights at urban road intersections are the same as the features disclosed in the methods of the above embodiments, and will not be repeated here.

[0081] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0082] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. All equivalent structural transformations made under the technical concept of this application using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included within the scope of patent protection of this application.

Claims

1. A multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections, characterized in that, The method includes: Acquire traffic phase time-series data and real-time traffic flow density data at the intersection, and perform fusion processing on the traffic phase time-series data and the real-time traffic flow density data to generate predicted lighting load power curve data; Based on the predicted lighting load power curve data, high-power pulse load segment data is identified, and pre-scheduling instruction data is generated according to the high-power pulse load segment data. Based on the pre-scheduling instruction data, the charging and discharging state of the high-power-density buffer unit in the heterogeneous energy storage unit is controlled, so that the high-power-density buffer unit enters a high-power-ready state. Within the time period corresponding to the high-power pulse load segment data, the real-time power demand data of the LED lighting module is obtained, the real-time power demand data is compared with the preset basic lighting power threshold, and dynamic power allocation instruction data is generated based on the comparison result. In response to the dynamic power allocation command data, the high power density buffer unit is scheduled to output the instantaneous peak power component, and the high energy density main energy storage unit is controlled to output the background power component, wherein the sum of the instantaneous peak power component and the background power component satisfies the real-time demand power data.

2. The multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections as described in claim 1, characterized in that, The steps of fusing the intersection traffic phase time-series data and the real-time vehicle flow density data to generate predicted lighting load power curve data include: The traffic phase timing data at the intersection is analyzed, and the data of each phase stage and its corresponding duration for at least one future signal cycle are determined according to the predetermined phase switching schedule of the traffic lights. Based on a pre-trained phase-power mapping model, the corresponding reference lighting power level data is determined for each phase stage data. Based on the pre-trained traffic flow-power correction model, the real-time traffic flow density data is processed to calculate the real-time lighting power correction coefficient data. The reference lighting power level data and the corresponding real-time lighting power correction coefficient data are applied to the preset basic lighting power data to obtain the predicted lighting load power curve data.

3. The multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections as described in claim 2, characterized in that, The steps for processing the real-time traffic density data and calculating the real-time lighting power correction coefficient data based on the pre-trained traffic flow-power correction model include: The real-time traffic density data at the current moment is combined with the historical traffic density data within at least one historical time window to form a traffic density data sequence. The traffic flow density data sequence is input into the pre-trained traffic flow-power correction model so that the pre-trained traffic flow-power correction model outputs the real-time lighting power correction coefficient data.

4. The multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections as described in claim 1, characterized in that, The steps of identifying high-power pulse load segment data based on the predicted lighting load power curve data and generating pre-scheduling instruction data based on the high-power pulse load segment data include: The predicted lighting load power curve data is processed by moving average filtering to obtain smoothed power curve data; The smoothed power curve data is subjected to point-by-point difference calculation to obtain the power change rate data; In the smoothed power curve data, find continuous data segments where the power value exceeds the first power threshold and the corresponding power change rate data exceeds the first change rate threshold, and use them as high-power pulse candidate segment data; When the duration of the high-power pulse candidate segment data exceeds the shortest pulse duration threshold, the high-power pulse candidate segment data is marked as the high-power pulse load segment data.

5. The multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections as described in claim 4, characterized in that, The method further includes: When at least two high-power pulse load segment data are identified, the interval data between the start time points of two adjacent high-power pulse load segment data is calculated. If the interval data is less than the preset minimum interval threshold, the data of the at least two high-power pulse load segments are merged to generate new high-power pulse load segment data, and the start time data, duration data and peak power data of the new high-power pulse load segment data are updated.

6. The multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections as described in claim 1, characterized in that, Based on the pre-scheduling command data, the step of controlling the charging and discharging state of the high-power-density buffer unit in the heterogeneous energy storage unit, so that the high-power-density buffer unit enters a high-power-ready state, includes: The pre-scheduling instruction data is parsed to obtain the power buffer start time data and preset advance time data corresponding to the high-power pulse load segment data; At a time when the power buffer start-up time data is reached but is greater than or equal to the preset advance time data, a preparation command is sent to the power conversion circuit connected to the high power density buffer unit. The preparation command includes preparation voltage setting value data, which is used to control the terminal voltage of the high power density buffer unit. In response to the execution of the preparatory command, the output status data of the high power density buffer unit is monitored in real time; Based on the output status data, determine whether the high power density buffer unit is in an effective ready state; If the high power density buffer unit is not in an effective ready state, the ready voltage setting value data is updated according to a preset adjustment strategy, and a new ready command containing the updated ready voltage setting value data is sent so that the high power density buffer unit reaches the high power ready state.

7. The multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections as described in claim 6, characterized in that, The step of determining whether the high power density buffer unit is in an effective ready state based on the output status data includes: Extract steady-state monitoring current data from the output status data; Calculate the deviation rate between the steady-state monitoring current data and the preset target current data; If the deviation rate data exceeds a preset deviation threshold, it is determined that the high power density buffer unit is not in an effective preparation state.

8. The multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections as described in claim 1, characterized in that, The steps of acquiring real-time power demand data of the LED lighting module within the time period corresponding to the high-power pulse load segment data, comparing the real-time power demand data with a preset basic lighting power threshold, and generating dynamic power allocation instruction data based on the comparison result include: Within the time period corresponding to the high-power pulse load segment data, the real-time power demand data of the LED lighting module is obtained at a fixed sampling period; Calculate the absolute value of the difference between the real-time demand power data obtained in the current sampling period and the historical demand power data obtained in the previous sampling period; The standard comparison process is executed, and dynamic power allocation instruction data is generated and output based on the comparison results between the real-time demand power data obtained in the current sampling period and the basic lighting power threshold. If the absolute value of the difference exceeds the preset power mutation threshold, the fast response mode is activated synchronously, and fast response factor data is generated based on the absolute value of the difference. Based on the fast response factor data, the power setting parameters in the dynamic power allocation command data are corrected to obtain the corrected dynamic power allocation command data, which is then overwritten and output.

9. The multi-level dynamic scheduling method for photovoltaic energy storage streetlights at urban road intersections as described in claim 8, characterized in that, The method further includes: In the fast response mode, the real-time demand power data obtained in subsequent sampling periods is continuously monitored; If the absolute value of the difference data is less than the preset power mutation threshold within a consecutive preset number of sampling periods, then generate mode exit instruction data. In response to the mode exit command data, the system exits the fast response mode and stops modifying the dynamic power allocation command data based on the fast response factor data.

10. A multi-level dynamic scheduling system for photovoltaic energy storage streetlights at urban road intersections, characterized in that, The photovoltaic energy storage street light multi-level dynamic scheduling system at urban road intersections includes: a memory, a processor, and a photovoltaic energy storage street light multi-level dynamic scheduling program for urban road intersections stored in the memory and executable on the processor. The photovoltaic energy storage street light multi-level dynamic scheduling program for urban road intersections is configured to implement the steps of the photovoltaic energy storage street light multi-level dynamic scheduling method for urban road intersections as described in any one of claims 1 to 9.

Images