Power distribution area voltage out-of-limit governance method and system based on automatic driving energy storage vehicle

Abstract

This invention discloses a method and system for managing voltage exceedances in distribution transformer areas based on autonomous driving energy storage vehicles, relating to the field of distribution network voltage control technology. The method includes acquiring transformer area data, energy storage vehicle data, and regional road network data; constructing a voltage prediction model based on historical data to map net load to transformer voltage; using the model to convert day-ahead net load prediction data into day-ahead voltage prediction data, and determining exceedances based on voltage assessment indicators to obtain exceedance information; based on the exceedance information, energy storage vehicle data, and road network data, constructing a joint optimization model with voltage compliance as a constraint and optimal operating cost as the objective, and solving for the day-ahead scheduling scheme for each vehicle; executing the day-ahead scheduling scheme to achieve voltage exceedance management. This invention achieves automated and economical management of voltage exceedances by enabling energy exchange between exceedance transformer areas using autonomous driving energy storage vehicles.

Description

Technical Field

[0001] This invention belongs to the field of distribution network voltage control technology, and particularly relates to a method and system for managing voltage over-limit in distribution substations based on autonomous driving energy storage vehicles. Background Technology

[0002] In power distribution networks, as the proportion of distributed photovoltaic (PV) systems connected to residential and industrial / commercial parks continues to increase, the problem of load imbalance on the low-voltage side of distribution transformers is becoming increasingly prominent. Specifically, during peak PV output periods at midday (i.e., when PV output exceeds local load demand), some distribution transformers experience power backflow into the grid due to PV output far exceeding local load demand, causing the low-voltage side voltage to exceed its upper limit. Conversely, during peak load periods in the evening and at night, the rapid increase in load on distribution transformers leads to significant voltage drops at the ends of branch lines, easily resulting in voltage exceeding the lower limit. These voltage exceeding issues directly impact power quality and have become one of the key factors restricting further integration of distributed PV and the safe operation of the power distribution network.

[0003] To address the aforementioned voltage exceedance issues, a common approach is to deploy stationary energy storage systems within the distribution transformer area. The basic principle is that energy storage absorbs excess energy during peak photovoltaic output periods and releases it during peak load periods, thereby mitigating source-load imbalances and reducing voltage deviations. However, once stationary energy storage equipment is built, its capacity, location, and operating parameters remain relatively fixed, making it difficult to flexibly relocate with seasonal changes, weather variations, or dynamic evolution of the source-load structure. When the spatial distribution and duration of voltage exceedance issues change significantly, the effectiveness of existing equipment is often limited, and the marginal cost of mitigation is high, making it difficult to achieve cost-effective and efficient routine operation.

[0004] With the continuous development of portable power technology and vehicle electric drive technology, mobile energy storage systems offer a new technological approach for voltage management in power distribution networks. Mobile energy storage can facilitate cross-regional energy exchange between transformer substations exceeding upper and lower limits, offering advantages such as mobile deployment, rapid access, and centralized maintenance. It can also achieve flexible adjustment without altering the existing power distribution structure. However, currently, mobile energy storage systems largely rely on manual driving and on-site operation. Task assignment, on-site and off-site coordination, and charging / discharging power settings all require manual intervention, resulting in high scheduling complexity and high labor costs, making it difficult to achieve continuous and economical routine operation across multiple transformer substations, time periods, and scenarios. Summary of the Invention

[0005] To address the aforementioned shortcomings in existing technologies, this invention provides a method and system for managing voltage overruns in distribution transformer substations based on autonomous driving energy storage vehicles. The aim is to effectively reduce the occurrence rate of voltage overruns in distribution transformer substations, improve voltage qualification rate and power supply quality, and balance dispatch economy and technological replicability without increasing investment in fixed equipment.

[0006] This invention solves the above-mentioned technical problems through the following technical solution: a method for managing voltage exceedance in distribution substations based on autonomous driving energy storage vehicles, comprising:

[0007] Acquire data on transformer substations, energy storage vehicles, and regional road networks within the target area; the transformer substation data includes at least the voltage assessment indicators, historical voltage data, historical net load data, and day-ahead net load forecast data for each substation; the energy storage vehicle data includes at least the current location, state of charge, and power parameters of each vehicle; the regional road network data includes at least the travel time information of energy storage vehicles between road nodes.

[0008] Based on the historical voltage data and historical net load data, a voltage prediction model is constructed to map the net load of the distribution area to the voltage of the distribution transformer.

[0009] According to the voltage prediction model, the day-ahead net load prediction data is converted into day-ahead voltage prediction data, and the day-ahead voltage prediction data is judged to exceed the limit based on the voltage assessment index to obtain the limit exceedance information; the limit exceedance information includes at least the limit exceedance area, the limit exceedance period, and the voltage regulation demand.

[0010] Based on the over-limit information, energy storage vehicle data, and regional road network data, a joint optimization model is constructed with voltage compliance as a constraint and optimal operating cost as the objective. The joint optimization model is solved to obtain the day-ahead scheduling scheme for each vehicle. The day-ahead scheduling scheme includes at least the travel route, service sequence, and power scheduling instructions for the vehicle in the transformer area.

[0011] The day-ahead scheduling scheme is implemented to address voltage over-limit issues in distribution transformer areas.

[0012] This invention acquires historical voltage and net load data to construct a load-voltage mapping relationship, converts day-ahead net load forecast data into day-ahead voltage forecast data, and then combines this with voltage assessment indicators to determine over-limit conditions. This allows for accurate identification of transformer areas exceeding upper and lower voltage limits, along with their corresponding over-limit periods and voltage regulation needs. This process eliminates the traditional reliance on experience-based judgment or simple threshold comparisons, enabling subsequent scheduling decisions to be based on quantitative forecasts, thus improving the targeting and reliability of governance.

[0013] The day-ahead scheduling scheme obtained by this invention can pair voltage-limit-above-limit and voltage-limit-below-limit transformer areas for energy mutual assistance, enabling energy storage vehicles to cycle between different transformer areas for charging and discharging, thus achieving spatiotemporal matching of source and load. Compared with stationary energy storage, energy storage vehicles can be flexibly deployed according to the spatial distribution and duration of voltage limits, adapting to the dynamic changes in seasons, weather, and source-load structure, overcoming the limitations of fixed equipment capacity and rigid location, and improving the adaptability and effectiveness of governance.

[0014] By combining day-ahead scheduling with autonomous energy storage vehicles and automated charging / discharging stations, tasks such as route navigation, charging / discharging access, and power closed-loop correction can be completed automatically without manual driving or on-site operation. This not only reduces the uncertainty caused by human intervention but also significantly reduces the scheduling manpower costs and operational complexity in multi-region and multi-time-period scenarios, making routine operation possible.

[0015] By constructing a joint optimization model with the goal of "optimal operating cost," the day-ahead dispatching scheme comprehensively considers multiple factors such as travel routes, service sequences, and charging / discharging power. This minimizes the total operating cost of the energy storage vehicle within the dispatching cycle while meeting voltage performance targets. This optimization process considers economic efficiency from the scheme formulation stage, ensuring that voltage management no longer relies solely on fixed equipment investment. Instead, it achieves a balance between voltage quality improvement and operational economy through the flexible dispatching of mobile energy storage, without adding new fixed equipment. This demonstrates good technical and economic viability and promising prospects for widespread application.

[0016] Furthermore, the voltage prediction model is constructed, including:

[0017] The historical voltage data and historical net load data are preprocessed to obtain preprocessed sample data;

[0018] Construct auxiliary features for modeling, the auxiliary features being used to characterize time attributes, environmental attributes and / or operating scenarios; group the sample data according to the auxiliary features, classifying data with the same or similar features into the same group;

[0019] For each subgroup, the voltage prediction model is determined based on the sample data within that subgroup, and then the hyperparameters of the voltage prediction model are determined; wherein, the distribution transformer voltage is expressed by the following formula:

[0020] ;

[0021] in, This represents the low-voltage side voltage of the i-th distribution area at time t, i.e., the distribution transformer voltage. This represents the net load of the i-th transformer area; Indicate auxiliary features; Represents the residual term; Indicates that the hyperparameters are Voltage prediction model; the hyperparameters The estimation is performed by minimizing the objective function, which is expressed as:

[0022] ;

[0023] in, This represents the hyperparameter estimates; Indicates sample weights; Represents the loss function; Represents the regularization coefficient; represents the regularization term; T represents the total number of sampling times within the target time period.

[0024] By constructing auxiliary features such as time period, season, holiday, and weather, and then grouping the samples according to these features and modeling them separately, the model can learn the mapping relationship between net load and voltage for different times, environments, and operating scenarios. This avoids the limitation of a single model being unable to take into account multiple operating conditions, thereby improving the adaptability of voltage prediction in different scenarios.

[0025] Introducing a regularization term during hyperparameter estimation can effectively control model complexity, prevent overfitting, and achieve a good balance between the model's fitting accuracy on training samples and its predictive generalization ability on new samples, thus ensuring the reliability of voltage prediction in practical applications.

[0026] Furthermore, based on the voltage assessment indicators, the day-ahead voltage forecast data is judged to exceed the limit, including:

[0027] Based on the voltage assessment indicators, set the upper voltage threshold and the lower voltage threshold, and set the upper limit margin and the lower limit margin;

[0028] For the day-ahead voltage forecast data, the following parameters for each transformer area are statistically analyzed according to the target time period:

[0029] Duration exceeding the upper limit: The cumulative duration of the day-ahead voltage forecast data ≥ the sum of the upper voltage limit threshold and the upper limit margin;

[0030] Duration beyond the lower limit: The cumulative duration of the difference between the current day voltage forecast data and the lower voltage limit threshold and the lower limit margin;

[0031] Percentage of samples exceeding the upper limit: The proportion of times when the current day voltage prediction data is greater than or equal to the sum of the upper limit threshold and the upper limit margin, out of the total number of sampling times;

[0032] Percentage of samples exceeding the lower limit: The proportion of times when the current day voltage prediction data is ≤ the difference between the lower voltage limit threshold and the lower limit margin out of the total number of sampling times;

[0033] By combining preset duration thresholds and sample proportion thresholds, the type of each transformer area is determined:

[0034] If the duration of a transformer area exceeding the upper limit is not less than the corresponding duration threshold, or the proportion of samples exceeding the upper limit is not less than the corresponding sample proportion threshold, it is determined to be an upper limit transformer area; if the duration of a transformer area exceeding the lower limit is not less than the corresponding duration threshold, or the proportion of samples exceeding the lower limit is not less than the corresponding sample proportion threshold, it is determined to be a lower limit transformer area; otherwise, it is determined to be a normal transformer area.

[0035] For transformer areas that exceed the upper or lower limit, an over-limit indication sequence is generated based on the day-ahead voltage prediction data; the over-limit indication sequence is used to mark whether a voltage over-limit has occurred at each time.

[0036] Connectivity segment extraction is performed on the over-limit indication sequence to obtain multiple over-limit intervals;

[0037] Within each over-limit interval, the voltage regulation requirement is calculated, and the over-limit information is output; the voltage regulation requirement represents the amount of voltage regulation required to restore the voltage within that interval to the assessment interval; the assessment interval consists of an upper voltage threshold and a lower voltage threshold.

[0038] By setting upper and lower voltage thresholds, as well as upper and lower margins, and statistically analyzing the cumulative duration and percentage of voltage exceeding the upper / lower limits, a comprehensive judgment can be made by combining the duration threshold and the percentage threshold. This effectively distinguishes between short-term voltage fluctuations and substantial voltage exceeding limits, avoiding misjudgments caused by transient interference, thereby improving the accuracy of transformer substation type identification.

[0039] Within each identified over-limit range, voltage regulation requirements are calculated, quantifying the voltage over-limit problem into specific regulation targets. This quantified output provides a direct basis for estimating the charging and discharging power and energy demand required by energy storage vehicles in subsequent optimized scheduling, making scheduling decisions more accurate and efficient.

[0040] By identifying over-limit transformer areas, extracting over-limit time periods, and outputting voltage regulation requirements, the subsequent joint optimization model can focus on the transformer areas and time periods that truly need regulation, avoiding unnecessary scheduling of normal transformer areas, thereby reducing the ineffective driving and charging / discharging losses of energy storage vehicles, and improving overall scheduling efficiency and regulation effectiveness.

[0041] Furthermore, the determination of the type of each station area also includes:

[0042] Hysteresis interval is used to process the judgment boundary; the hysteresis interval is a voltage range defined by a first boundary threshold and a second boundary threshold, wherein the first boundary threshold is the judgment boundary for entering the over-limit state from the normal state, and the second boundary threshold is the judgment boundary for recovering from the over-limit state to the normal state, and the first boundary threshold and the second boundary threshold are different.

[0043] For the voltage upper limit side, the first boundary threshold is the voltage upper limit threshold, and the second boundary threshold is the voltage upper limit threshold minus a preset hysteresis width. The processing is specifically as follows:

[0044] When the voltage rises from the assessment range and crosses the first boundary threshold, it is determined that the voltage exceeds the upper limit;

[0045] It is determined that the voltage has returned to normal only when it drops from the upper limit and crosses the second boundary threshold.

[0046] For the lower voltage limit side, the first boundary threshold is the lower voltage limit threshold, and the second boundary threshold is the lower voltage limit threshold plus a preset hysteresis width. The specific processing is as follows:

[0047] When the voltage drops from the assessment range and crosses the first boundary threshold, it is determined that the voltage has exceeded the lower limit.

[0048] It is determined that the voltage has returned to normal only when it rises from the lower limit and crosses the second boundary threshold.

[0049] By setting different entry thresholds (first boundary threshold) and exit thresholds (second boundary threshold) for the upper and lower voltage limits respectively, a hysteresis range is formed. When the voltage fluctuates near the threshold, as long as it does not cross the exit threshold, the distribution area type determination result will remain unchanged. This mechanism effectively avoids the distribution area type from repeatedly switching between "over-limit" and "normal" due to instantaneous voltage fluctuations, making the determination result more stable and reliable. Stable distribution area type determination results avoid invalid dispatch instructions triggered by frequent switching. Energy storage vehicles do not need to frequently adjust their dispatch plans due to short-term voltage fluctuations, reducing unnecessary driving, access, and charging / discharging operations, thereby reducing system operating costs and improving overall dispatch efficiency.

[0050] Furthermore, after obtaining multiple over-limit intervals, the process also includes:

[0051] Adjacent out-of-limit intervals are merged; the merging process includes:

[0052] Calculate the time interval between two adjacent over-limit intervals. If the time interval is less than a preset merging threshold, then merge the two over-limit intervals into one over-limit interval.

[0053] By merging adjacent out-of-limit intervals with a time interval less than a preset merging threshold into a single continuous out-of-limit interval, the originally continuous governance needs are prevented from being split into multiple independent time periods due to a brief voltage drop. Energy storage vehicles no longer need to frequently travel between multiple short time periods, reducing the number of trips and access / disconnect operations, thereby lowering total mileage, energy consumption, and equipment wear and tear.

[0054] The merged over-limit zones have a longer continuous governance time, which facilitates the continuous and stable charging and discharging services of energy storage vehicles in a single distribution area, avoiding excessive energy loss and response delays caused by multiple start-stop cycles. This approach makes subsequent charging and discharging power estimation and service time scheduling more reasonable, improving the executability and governance efficiency of the overall scheduling scheme.

[0055] Furthermore, based on the aforementioned over-limit information, energy storage vehicle data, and regional road network data, a joint optimization model is constructed with voltage compliance as a constraint and optimal operating cost as the objective, including:

[0056] For each over-limit transformer area, based on its over-limit period and voltage regulation requirements, determine the theoretical power regulation amount required to restore its voltage to meet the voltage assessment indicators, and thereby determine the power regulation direction for that transformer area.

[0057] Based on the power adjustment direction of each transformer area, cross-limit transformer areas with opposite power adjustment directions are paired with each other, or cross-limit transformer areas that need adjustment are paired with normal transformer areas that do not need adjustment, to generate candidate transformer area mutual assistance pairs.

[0058] Based on the current location, state of charge, and travel time information of each vehicle, the feasibility of each candidate transformer station mutual assistance pair is screened to obtain feasible pairing relationships; the conditions for feasibility screening include: the vehicle's accessibility and the vehicle's state of charge not being lower than a preset state of charge limit upon arrival.

[0059] The energy demand is calculated based on the theoretical power adjustment, the executable power is determined based on the vehicle's power parameters and power output capability, and then the service time demand is calculated based on the energy demand and the executable power.

[0060] The joint optimization model is constructed, which includes a comprehensive objective function J and constraints. The expression for the comprehensive objective function J is:

[0061] ;

[0062] in, This represents the total travel time of vehicle k calculated based on travel time information. This represents the driving energy consumption of vehicle k, calculated based on energy consumption per unit mileage and distance traveled. B represents the energy loss during charging and discharging calculated based on the vehicle's charging and discharging efficiency; R represents the vehicle load balance, used to measure the balance of tasks undertaken by each vehicle within the scheduling cycle; and R represents the reserve margin of the vehicle's state of charge. , , , This represents the reference constant for normalization of each term; , , , The parameters representing the weights of each item;

[0063] The constraints include voltage constraints, feasible pairing constraints, service energy constraints, service time constraints, energy storage state of charge constraints, charge and discharge power constraints, accessibility constraints, and energy reservation constraints.

[0064] Candidate transformer area mutual aid pairs are generated by pairing transformer areas with opposite power regulation directions, or between transformer areas requiring regulation and normal transformer areas that do not require regulation. This pairing mechanism ensures that the energy flow direction matches the voltage over-limit type, enabling energy storage vehicles to absorb power in transformer areas exceeding the upper voltage limit and release power in transformer areas exceeding the lower voltage limit, achieving precise energy mutual aid between transformer areas and avoiding ineffective or reverse scheduling. The feasibility screening mechanism for candidate transformer area mutual aid pairs ensures that the generated scheduling scheme is executable in actual operation, avoiding scheduling failures due to insufficient travel time, insufficient energy, or time window conflicts, thus improving the feasibility of the scheme.

[0065] By calculating energy demand based on theoretical power regulation, determining executable power based on vehicle power parameters and output capabilities, and then calculating service time requirements based on energy demand and executable power, the abstract voltage limit violation problem is quantified into the amount of energy compensation that the energy storage vehicle can perform (energy demand), the physically achievable maximum output (executable power), and the time required to complete compensation (service time requirements). This quantification process provides clear physical boundaries and constraint parameters for subsequent power dispatch command generation and service timing arrangements, enabling dispatch decisions to accurately match voltage management needs with the actual capabilities of the energy storage vehicle, avoiding management failures due to insufficient energy or power limitations, thereby improving the executability and efficiency of the dispatch scheme.

[0066] The comprehensive objective function integrates multiple objectives, including driving time, driving energy consumption, charging and discharging energy loss, load balancing, and SOC reserve margin, and achieves multi-objective collaborative optimization through normalization and weighting parameters. Constraints cover multiple dimensions such as voltage, energy, power, time window, and state of charge, ensuring that the solution simultaneously meets voltage assessment indicators and operational feasibility requirements. This multi-objective optimization framework enables energy storage vehicle scheduling to not only focus on voltage management effectiveness but also consider operational economy, equipment balancing, and energy security margin, achieving a balance between voltage quality improvement and operating cost control.

[0067] Further, executing the day-ahead scheduling scheme includes:

[0068] Control the energy storage vehicle to travel to the target distribution area and connect according to the day-ahead dispatch plan;

[0069] Real-time data acquisition of the target transformer area's voltage and the vehicle's real-time state of charge;

[0070] Based on the deviation between the real-time voltage and the voltage assessment index, and in conjunction with the vehicle's power parameters, the power dispatch command is adjusted in real time to generate the actual power command.

[0071] When a triggering event is detected that requires adjustment of the remaining schedule, the daytime scheduling plan for the remaining period is dynamically updated based on the vehicle's current location, real-time state of charge, power parameters, and remaining over-limit information, and a new scheduling plan is generated.

[0072] The above real-time adjustment and dynamic update process is executed iteratively until the voltage of each transformer area meets the voltage assessment index within the target time period.

[0073] By controlling energy storage vehicles to automatically travel to the target distribution area and connect according to the day-ahead dispatch plan, and combining real-time collected voltage and state-of-charge data, power dispatch instructions are adjusted in real time, forming a closed-loop control mechanism of "planning → execution → feedback → correction". This process requires no manual intervention, realizing the automated connection from dispatch plan formulation to execution, significantly reducing manual operation costs and on-site response delays.

[0074] During real-time execution, the power dispatch command is adjusted in real time based on the deviation between the real-time voltage and the voltage assessment index, enabling the energy storage vehicle to dynamically correct its output according to the actual voltage fluctuations. This overcomes the possible deviations in the day-ahead dispatch scheme based on forecast data and ensures that even if the forecast is inaccurate or the actual operating conditions fluctuate, the voltage in the distribution area can still be effectively maintained within the normal range.

[0075] When a triggering event is detected that requires adjustment to the remaining schedule (such as excessive voltage deviation, abnormal travel time, or insufficient service energy), the scheduling plan for the remaining time period can be dynamically updated based on the current state and remaining over-limit information. This rolling replanning mechanism enables the system to cope with uncertainties such as changes in road conditions, abnormal equipment status, and sudden load changes, avoiding overall governance failure due to a single fault or deviation, and improving the system's reliability and adaptability.

[0076] By iteratively executing real-time adjustments and dynamic updates until the voltage of each transformer substation meets the voltage assessment targets within the target time period, a cyclical control process of "execution → monitoring → adjustment → re-execution" is formed. This iterative mechanism ensures that regardless of whether there are deviations in the initial plan or any disturbances that occur during execution, the voltage of each transformer substation will ultimately meet the standards, guaranteeing the attainability and stability of the governance effect.

[0077] Furthermore, the real-time adjustment of the power scheduling command includes:

[0078] Obtain the power dispatching instruction for the vehicle in the current target area from the daytime dispatching scheme;

[0079] Determine the target voltage based on the type of task the vehicle is currently performing:

[0080] For charging tasks, the target voltage is the difference between the upper voltage threshold and the preset upper limit margin; for discharging tasks, the target voltage is the sum of the lower voltage threshold and the preset lower limit margin.

[0081] Based on the deviation between the target voltage and the real-time voltage, the power value at the current moment in the power scheduling command is proportionally corrected to generate the actual power command.

[0082] The actual power command is limited to the range of the vehicle's power parameters by a limiting operator, resulting in a corrected current actual power command.

[0083] By obtaining power dispatch instructions from the day-ahead dispatch plan and making proportional corrections based on real-time voltage deviations, the actual output of the energy storage vehicle can be dynamically adjusted according to real-time changes in the voltage of the distribution area. This mechanism effectively compensates for the discrepancy between day-ahead forecasts and actual operation, ensuring that even in cases of inaccurate forecasts or load fluctuations, the output of the energy storage vehicle can still accurately respond to voltage management needs.

[0084] By using a limiting operator, the actual power command is restricted to the vehicle's power parameters, ensuring that the generated command does not exceed the vehicle's actual power limits and climbing ability. This constraint guarantees the feasibility of real-time power commands at the execution level, preventing execution failures or equipment damage due to commands exceeding the vehicle's capabilities.

[0085] Target voltages are set according to the charging or discharging task type, and proportional corrections are made based on the deviation between the target voltage and the real-time voltage. This control strategy enables the energy storage vehicle to actively absorb power when the voltage exceeds the upper limit and actively inject power when the voltage exceeds the lower limit. The correction strength increases with the increase of the deviation, which can smoothly guide the voltage to the qualified range.

[0086] The real-time adjustment process is entirely automated based on preset control logic, requiring no manual monitoring or operation. Combined with autonomous driving energy storage vehicles and automated charging and discharging stations, it achieves fully automated closed-loop control from planned commands to real-time output, significantly reducing labor costs and operational complexity during operation.

[0087] Furthermore, the triggering event includes any of the following:

[0088] The deviation between the real-time voltage and the target voltage exceeds a preset voltage deviation threshold; the target voltage is determined based on the type of task currently being performed by the vehicle.

[0089] If the travel time, vehicle access status, or vehicle status calculated based on the travel time information is abnormal, the service sequence in the daytime scheduling scheme will be unreachable.

[0090] Before the planned departure time, the accumulated effective service energy is less than the energy requirement, or the accumulated effective service time is less than the service time requirement;

[0091] The effective service energy is the integral of the actual power command during the over-limit period, and the effective service time is the cumulative duration after the vehicle accesses the transformer area.

[0092] By setting three clearly defined trigger events—real-time voltage deviation exceeding a threshold, service unavailability due to abnormal travel time or status, and insufficient service energy or service time—a multi-dimensional, multi-scenario real-time monitoring system has been constructed. When actual operation deviates from the plan, the system can promptly detect and trigger replanning, avoiding governance failures due to prediction deviations, changes in road conditions, or equipment malfunctions, significantly improving the system's fault tolerance and operational reliability to uncertainties.

[0093] All three trigger events are based on comparisons between real-time data (real-time voltage, vehicle location, state of charge, connection status, etc.) and preset thresholds, eliminating the need for manual judgment and intervention. The system can autonomously identify abnormal situations and automatically initiate rolling replanning, forming an automated closed loop of "monitoring → triggering → adjustment," further reducing reliance on manual monitoring and intervention.

[0094] The "accumulated service insufficiency before the planned departure time" mechanism in the trigger conditions enables the system to initiate replanning in advance when it anticipates that a task cannot be completed before it has finished, rather than passively responding after the task fails. This preventative adjustment capability effectively avoids governance failures due to insufficient energy or time, improving the success rate of the overall scheduling scheme.

[0095] Furthermore, the day-ahead scheduling plan for the remaining time period is dynamically updated, including:

[0096] Calculate the remaining service demand for the current target area, whereby the remaining service demand includes the remaining energy demand and the remaining service time demand.

[0097] Using the current moment as the starting moment for dynamic updates, and combining the vehicle's current location, real-time state of charge, information on unserved substations, and updated road network travel time information, feasible pairing relationships are regenerated, executable power, energy demand, and service time demand are re-estimated, and the joint optimization model is re-solved to obtain the revised day-ahead scheduling scheme.

[0098] Based on the revised day-ahead scheduling plan, update the vehicle's subsequent travel routes, service sequence, and power scheduling instructions.

[0099] By calculating the remaining service demand for the current target area and using the current moment as the starting point for dynamic updates, subsequent optimizations can be precisely corrected based on the actual service volume already completed. This mechanism avoids redundant calculations requiring replanning from scratch, achieving seamless integration of the old and new scheduling schemes in terms of time and tasks, and improving the efficiency and accuracy of replanning.

[0100] During the dynamic update process, the system combines the vehicle's current location, real-time charge status, information on unserved substations, and updated road network travel time information to ensure that the regenerated scheduling plan fully reflects the vehicle's actual location, energy status, remaining tasks, and the latest road conditions. This feature ensures that the adjusted plan is highly targeted and executable, avoiding secondary deviations caused by outdated information.

[0101] Based on the revised day-ahead dispatching scheme, the subsequent travel routes, service sequences, and power dispatching instructions of the vehicles are updated, enabling energy storage vehicles to smoothly transition to the updated service sequence after completing their current tasks. This update mechanism ensures that vehicles maintain clear and feasible service arrangements throughout the entire dispatching cycle, avoiding task interruptions or execution chaos caused by scheme adjustments.

[0102] Based on the same concept, the present invention also provides a distribution area voltage over-limit management system based on an autonomous driving energy storage vehicle, including a memory, a processor, and a computer program or instructions stored in the memory. The processor executes the computer program or instructions to implement the distribution area voltage over-limit management method based on an autonomous driving energy storage vehicle as described above. Attached Figure Description

[0103] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only one embodiment of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0104] Figure 1 This is a flowchart of the distribution area voltage over-limit management method in an embodiment of the present invention. Detailed Implementation

[0105] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0106] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0107] Example 1

[0108] This invention provides a method for managing voltage exceedances in distribution substations based on autonomous driving energy storage vehicles. This method constructs a load-voltage mapping relationship (i.e., a voltage prediction model), generates voltage prediction data using day-ahead net load prediction data, and then identifies the substations exceeding the limit, their overdue periods, and voltage regulation needs. Based on this, a joint optimization model is constructed with voltage compliance as a constraint and optimal operating cost as the objective, to obtain the day-ahead scheduling scheme for each vehicle. During the execution phase, effective management of substation voltage is achieved through real-time adjustments and rolling replanning.

[0109] This embodiment takes a regional power distribution network as an example. This region includes multiple distribution substations, each equipped with several automated energy storage vehicles. Each substation is equipped with an automated charging / discharging station, and the energy storage vehicles and charging / discharging stations exchange data in real time via a 4G / 5G communication link. Figure 1 As shown, the method for managing voltage exceedance in this distribution transformer area includes the following steps:

[0110] Step S1: Obtain data on transformer substations, energy storage vehicles, and regional road networks within the target area.

[0111] The data for the distribution area includes:

[0112] Each transformer substation is equipped with latitude and longitude coordinates to determine its geographical location and subsequent route planning;

[0113] Voltage performance indicators for each transformer substation, including upper voltage thresholds. and voltage lower limit threshold In this embodiment, a voltage upper limit threshold is set with reference to the "Regulations on Power Supply Voltage Management of State Grid Corporation of China". Voltage lower limit threshold Voltage deviation threshold ;

[0114] Historical voltage data of the low-voltage side of the distribution transformer, sampling time step It lasts for 15 minutes;

[0115] Historical net load data for the distribution area (defined as the difference between distributed generation output and electricity load), sampling time step. The interval is 15 minutes, aligned with the historical voltage data time.

[0116] In addition, there is the day-ahead net load forecast data, with a forecast time step of 15 minutes and a total of 96 forecast points (corresponding to 24 hours), which are used to generate the day-ahead voltage forecast data.

[0117] Energy storage vehicle data includes:

[0118] The current location (latitude and longitude coordinates) of each vehicle is used for route planning and scheduling decisions;

[0119] The state of charge (SOC) of each vehicle reflects the remaining battery power in real time.

[0120] SOC allows for upper and lower limits of operation; in this embodiment, the lower limit of SOC is set to... The upper limit of SOC is ;

[0121] Power parameters, also known as power charging and discharging parameters, specifically include rated charging power. Rated discharge power Climbing speed Charging efficiency Discharge efficiency ;

[0122] and target SOC reserved value It is used for energy reservation constraints.

[0123] Regional road network data includes travel time information for energy storage vehicles between road nodes. This communication time information is represented by a travel time matrix, where each element represents the travel time between any two distribution areas (road nodes). In this embodiment, the travel time matrix is ​​updated every 15 minutes based on historical traffic flow data to reflect real-time traffic conditions. Vehicle access preparation time. This is used to calculate the arrival and completion time of access. The charging and discharging service end time for each distribution station area. The time will be uniformly set to 3:00 AM the following day.

[0124] The above data is collected and updated in real time through the automated monitoring system of the distribution area, the on-board terminal of the energy storage vehicle, and the traffic information platform, providing a data foundation for subsequent model building, limit violation judgment, optimized scheduling, and execution feedback.

[0125] Step S2: Based on historical voltage data and historical net load data, construct a voltage prediction model to map the net load of the distribution area to the voltage of the distribution transformer.

[0126] In this embodiment, the specific construction process of the voltage prediction model (i.e., the load-voltage mapping relationship) includes:

[0127] Step S2.1: Preprocess the historical voltage data of the low-voltage side of the distribution transformer and the historical net load data of the transformer substation to obtain preprocessed sample data. The preprocessing operations include:

[0128] Data cleaning: Remove obviously abnormal data records, such as data points where the voltage value exceeds the physical reasonable range or the net load value significantly exceeds the capacity range of the transformer area.

[0129] Noise reduction: Moving average filtering or wavelet denoising methods are used to eliminate measurement noise and transient interference, while preserving the true trend of voltage and net load changes.

[0130] Missing data completion: For missing data caused by communication interruption or measurement failure, linear interpolation or the mean of historical data from the same period is used to fill in the missing data to ensure the continuity of the time series.

[0131] Time alignment: Historical voltage data and historical load data are aligned with a unified timestamp to ensure that at the same time, the historical voltage data and historical load data correspond to the same sampling point.

[0132] Step S2.2: Auxiliary feature construction and sample clustering.

[0133] Auxiliary features are constructed for modeling, representing temporal attributes, environmental attributes, and / or operational scenarios. In this embodiment, the constructed auxiliary features include:

[0134] Time attributes: Time period (dividing a 24-hour day into 96 15-minute time periods, or into peak, off-peak, and valley time periods), whether it is a weekday / holiday, and what day of the week it is;

[0135] Environmental attributes: air temperature (°C), irradiance (W / m²) 2 Meteorological elements such as wind speed (m / s) and humidity (%);

[0136] Operating scenarios: photovoltaic power output level (which can be obtained through historical data clustering), load level, etc.

[0137] After the above preprocessing and the addition of auxiliary features, a sample dataset for modeling is obtained. Each sample contains the transformer area number i, time t, net load value, voltage value, and auxiliary features.

[0138] The sample data is clustered based on auxiliary features, grouping data with the same or similar auxiliary features into the same cluster. The purpose of clustering is to ensure that samples within each cluster have similar source payload characteristics, thereby establishing a more targeted mapping relationship. This embodiment employs the following clustering strategy:

[0139] Grouped by season: Spring (March to May), Summer (June to August), Autumn (September to November), Winter (December to February);

[0140] Grouped by time period: Night (0:00~6:00), Morning (6:00~12:00), Afternoon (12:00~18:00), Evening (18:00~24:00);

[0141] Grouped by weather type: sunny, cloudy, overcast, rainy.

[0142] In actual clustering, a single dimension or a combination of dimensions can be selected for clustering based on the amount of data and actual needs. For example, samples with "summer + afternoon + sunny day" can be divided into a cluster, and the samples in this cluster have the typical characteristics of high temperature, strong sunlight, and high photovoltaic output.

[0143] Step S2.3: For each cluster, determine the voltage prediction model based on the sample data within that cluster, and then determine the hyperparameters of the voltage prediction model.

[0144] This embodiment uses Gradient Boosting Decision Tree (GBDT) as the base model, training a GBDT model for each cluster to obtain the voltage prediction model for that cluster. For each sample within a cluster, at discrete time t, the transformer voltage of transformer area i is... It can be expressed as follows:

[0145] (1)

[0146] in, This represents the low-voltage side voltage of the i-th distribution area at time t, i.e., the distribution transformer voltage (derived from historical voltage data). This represents the net load of the i-th transformer area (derived from historical net load data, measured value). It indicates auxiliary features, including time period, season, holiday markers, meteorological elements, etc. This represents the residual term, used to characterize random fluctuations that the model cannot fully explain; Indicates that the hyperparameters are The voltage prediction model, i.e., the trained GBDT model.

[0147] Model hyperparameters Estimation is performed by minimizing the objective function. In this embodiment, the objective function adopts a regularized empirical risk minimization form:

[0148] (2)

[0149] in, This represents the hyperparameter estimates; This represents the sample weight, which can be set according to factors such as the sample's time proximity and importance. In this embodiment, the default value is 1. The loss function is used to measure the difference between the model's predicted value and the true value. This embodiment uses the Huber loss function, which has the advantages of both mean squared error and absolute error and has good robustness to outliers. This represents the regularization coefficient, used to control model complexity and prevent overfitting. Its value can be determined through cross-validation. ; This represents the regularization term; in this embodiment, L2 norm regularization is used. T represents the total number of sampling times within the target time period.

[0150] The specific process of hyperparameter estimation is as follows:

[0151] The samples within each cluster are divided chronologically, with the first 80% used as the training set and the last 20% as the validation set. A grid search is performed on candidate hyperparameter combinations (such as the number of trees, maximum depth, learning rate, and regularization coefficient λ). K-fold cross-validation (k=5) is performed on the training set, and the average loss for each hyperparameter combination is calculated. The hyperparameter combination with the minimum loss on the validation set is selected as the validation set. Using all training samples with optimal hyperparameters Train the final GBDT model.

[0152] For each cluster, the above process is repeated to obtain the voltage prediction model corresponding to that cluster. In actual prediction, the cluster is first determined based on auxiliary features (season, time period, weather, etc.) of the time to be predicted. Then, the model for that cluster is called to perform the prediction, obtaining the predicted voltage value. .

[0153] Through the above-mentioned group modeling and parameter determination process, a voltage prediction model that can accurately reflect the mapping relationship between the net load of each transformer area and the voltage of the distribution transformer is finally obtained, providing a basis for subsequent voltage over-limit judgment.

[0154] Step S3: Based on the voltage prediction model, convert the day-ahead net load prediction data into day-ahead voltage prediction data, and determine the over-limit of the day-ahead voltage prediction data based on the voltage assessment index to obtain over-limit information.

[0155] In this embodiment, the specific implementation process of step S3 includes:

[0156] Step S3.1: Based on the voltage prediction model, convert the day-ahead net load prediction data into day-ahead voltage prediction data.

[0157] Net load forecast data as of today Input the voltage prediction model obtained in step S2 to calculate the day-ahead voltage prediction data of the low-voltage side of the distribution transformer in each area. :

[0158] (3)

[0159] in, Auxiliary features representing the current prediction time; This represents the voltage prediction model for the group corresponding to transformer area i.

[0160] Step S3.2: Set the upper voltage threshold and lower voltage threshold according to the voltage assessment indicators, and set the upper limit margin and lower limit margin.

[0161] In this embodiment, a voltage upper limit threshold is set with reference to the "State Grid Corporation of China Power Supply Voltage Management Regulations". Voltage lower limit threshold , upper limit margin Lower margin The assessment range consists of an upper voltage threshold and a lower voltage threshold. The control range consists of the assessment interval and the margin. .

[0162] Step S3.3: Analyze the day-ahead voltage forecast data The following parameters for each transformer substation are calculated according to the target time period (6:00 to 22:00 daily as the treatment period in this embodiment):

[0163] Exceeding the maximum duration The cumulative duration of the current voltage forecast data being greater than or equal to the sum of the upper voltage threshold and the upper voltage margin, i.e.:

[0164] (4)

[0165] Duration beyond the lower limit The cumulative duration during which the current voltage forecast data is ≤ the difference between the lower voltage threshold and the lower margin, i.e.:

[0166] (5)

[0167] Percentage of samples exceeding the upper limit The percentage of times when the predicted voltage data is greater than or equal to the sum of the upper voltage threshold and the upper voltage margin, out of the total number of sampling times.

[0168] (6)

[0169] percentage of samples exceeding the lower limit The proportion of the number of times when the current voltage prediction data is ≤ the difference between the lower voltage threshold and the lower margin out of the total number of sampling times, i.e.:

[0170] (7)

[0171] in, This indicates that the value is 1 when the condition inside the parentheses is true, and 0 otherwise.

[0172] Step S3.4: Combine the preset duration threshold and sample proportion threshold to determine the type of each station area.

[0173] In this embodiment, a duration threshold is set. Sample proportion threshold The duration threshold used for judging whether an item exceeds the upper or lower limit can be the same or different. Similarly, the sample percentage threshold used for judging whether an item exceeds the upper or lower limit can be the same or different.

[0174] The specific judgment rules are as follows:

[0175] If the duration of region i exceeds the upper limit or the proportion of samples exceeding the upper limit If the duration of the lower limit exceeding the limit for transformer area i is considered as exceeding the upper limit, then the transformer area is determined to be an area exceeding the upper limit. or the proportion of samples exceeding the lower limit If the condition is met, the transformer area is determined to be a transformer area that exceeds the lower limit; otherwise, it is determined to be a normal transformer area.

[0176] To avoid frequent switching of transformer substation types due to voltage fluctuations near the threshold, this embodiment uses a hysteresis interval to process the judgment boundary. The hysteresis interval is the voltage range defined by a first boundary threshold and a second boundary threshold, where the first boundary threshold is the judgment boundary for entering the over-limit state from the normal state, and the second boundary threshold is the judgment boundary for recovering from the over-limit state to the normal state, and the first boundary threshold and the second boundary threshold are different.

[0177] For the upper voltage limit side:

[0178] The first boundary threshold is the upper voltage threshold. The second boundary threshold is the upper voltage threshold minus the preset hysteresis width. In this embodiment, the hysteresis width is taken as 0.02pu.

[0179] When the voltage rises from the assessment range and crosses the first boundary threshold, it is determined that the voltage has exceeded the upper limit; only when the voltage falls from the upper limit and crosses the second boundary threshold is it determined that it has returned to normal.

[0180] For the lower voltage limit side:

[0181] The first boundary threshold is the lower voltage threshold. The second boundary threshold is the lower voltage threshold plus a preset hysteresis width.

[0182] When the voltage drops from the assessment range and crosses the first boundary threshold, it is determined that the voltage has exceeded the lower limit; only when the voltage rises from the lower limit and crosses the second boundary threshold is it determined that it has returned to normal.

[0183] By processing the hysteresis interval as described above, the problem of frequent switching of transformer substation type when the voltage fluctuates near the threshold is effectively avoided.

[0184] Step S3.5: Extraction and merging of out-of-limit intervals.

[0185] For transformer substations determined to have exceeded the upper or lower voltage limits, an over-limit indication sequence is generated based on their day-ahead voltage prediction data. The over-limit indication sequence is a binary sequence used to mark whether a voltage over-limit has occurred at each time point.

[0186] Define the voltage over-limit indicator sequence: ;

[0187] Define the voltage lower limit indication sequence: .

[0188] Connectivity segment extraction is performed on the out-of-limit indicator sequence to obtain multiple initial out-of-limit intervals. Connectivity segment extraction refers to identifying time segments with consecutive values ​​of 1 as independent out-of-limit intervals.

[0189] Adjacent out-of-limit intervals are merged: the time interval between two adjacent out-of-limit intervals (i.e., the difference between the end time of the previous interval and the start time of the next interval) is calculated. If the time interval is less than a preset merging threshold, the two out-of-limit intervals are merged into one out-of-limit interval. In this embodiment, the merging threshold is set to 15 minutes.

[0190] After merging, the final over-limit intervals are obtained. For each over-limit transformer area, the set of over-limit intervals is obtained. , and Let these represent the start and end times of the m-th upper limit crossing interval, respectively; for the lower limit crossing interval, we obtain the set of lower limit crossing intervals. , and These represent the start and end times of the nth interval that crosses the lower limit, respectively.

[0191] Step S3.6: Calculate the voltage regulation requirement within each over-limit range and output the over-limit information.

[0192] Voltage regulation requirement indicates the amount of voltage regulation required to restore the voltage within the specified range to the control range.

[0193] Voltage regulation requirements at each time t within each over-limit interval Calculate using the following formula:

[0194] (8)

[0195] The above calculation results are output in a structured format as input for subsequent optimized scheduling. The output over-limit information includes: the over-limit transformer area and its type (over-limit upper limit area or over-limit lower limit area); the over-limit time period corresponding to each over-limit transformer area (i.e., the extracted and merged over-limit interval); and the voltage regulation requirement at each over-limit time. .

[0196] This information on exceeding limits provides key input parameters for building the joint optimization model, estimating executable power, energy requirements, and service time requirements in step S4.

[0197] Step S4: Based on over-limit information, energy storage vehicle data, and regional road network data, construct a joint optimization model with voltage compliance as a constraint and optimal operating cost as the objective; solve the joint optimization model to obtain the day-ahead scheduling scheme for each vehicle.

[0198] The current dispatching scheme includes at least the travel route, service sequence, and power dispatching instructions for vehicles in the transformer substation area. In this embodiment, the specific implementation process of step S4 includes:

[0199] Step S4.1: For each over-limit transformer area, based on its over-limit period and voltage regulation requirements, determine the theoretical power regulation amount required to restore its voltage to meet the voltage assessment indicators, and thereby determine the power regulation direction of the transformer area.

[0200] The over-limit interval of station area i For example, at each time point within that interval The voltage sensitivity to power is obtained by using local linearization of the voltage prediction model. :

[0201] (9)

[0202] in, This represents the net load of the i-th transformer area; This is a small power disturbance (1kW in this embodiment).

[0203] Theoretical power regulation for:

[0204] (10)

[0205] in, This indicates that the energy storage vehicle needs to absorb active power from the transformer substation. This indicates that the energy storage vehicle is injecting active power into the transformer area. If If the value is too small or the mapping relationship contains a non-monotonic segment, then the solution can be obtained using the bisection method or Newton's iteration. This makes the adjusted electricity Falling into the control zone and Minimum.

[0206] Based on theoretical power regulation Define the direction of power demand :

[0207] This indicates that power needs to be injected ( This means that the energy storage vehicle needs to discharge to inject power into the power distribution area. Indicates the power required to be absorbed ( This means that the energy storage vehicle needs to absorb power from the power distribution area for charging. This indicates that no adjustment is needed.

[0208] Step S4.2: Based on the power adjustment direction of each transformer area, pair up transformer areas with opposite power adjustment directions that exceed the limit, or transformer areas that need adjustment and normal transformer areas that do not need adjustment, to generate candidate transformer area mutual assistance pairs.

[0209] Based on the power regulation direction of each transformer substation, the constructed candidate transformer substation mutual assistance pair set includes the following two types of pairings:

[0210] The first type of pairing consists of a voltage-above-the-upper limit transformer area and a voltage-above-the-lower limit transformer area or a normal transformer area, and the power demand direction of the voltage-above-the-upper limit transformer area is the direction of power absorption ( The power demand direction of the lower limit area or normal area is either requiring power injection or not requiring adjustment. );

[0211] The second type of pairing consists of a voltage lower limit area and a voltage upper limit area or a normal area, and the power demand direction of the voltage lower limit area is the direction of the power to be injected. The power demand direction of the above-limit transformer area or the normal transformer area is either that it needs to absorb power or does not need to be adjusted. ).

[0212] Step S4.3: Based on the current location, state of charge (SOC) of each vehicle, and travel time information from the regional road network data, perform feasibility screening on each candidate transformer substation mutual assistance pair to obtain feasible pairing relationships. The feasibility screening criteria include:

[0213] Driving accessibility: The time when a vehicle travels from its current location i to the target location j and completes the connection. No later than the preset end time of service in the district. . The calculation formula is:

[0214] (11)

[0215] in, Let k be the planned departure time of vehicle k in area i. This represents the travel time from area i to area j (obtained from the road network travel time matrix). Preparation time for access (2 minutes in this embodiment).

[0216] State of charge constraint: The state of charge (SOC) of the vehicle when it reaches the target area j shall not be lower than the preset lower limit of SOC. .

[0217] After feasibility screening, a set of feasible pairing relationships is obtained, which are feasible combinations that allow a certain vehicle to provide services to a certain transformer station within a certain over-limit range.

[0218] Step S4.4: For each over-limit area, calculate the energy demand based on the theoretical power adjustment amount, determine the executable power based on the vehicle's power parameters and power output capability, and then calculate the service time demand based on the energy demand and executable power.

[0219] In the over-limit zone Internal energy demand The integral of the absolute value of the theoretical power regulation:

[0220] (12)

[0221] Define an upper bound on the available power of vehicle k at time t:

[0222] (13)

[0223] in, The maximum charging and discharging power of vehicle k includes the rated charging power and the rated discharging power (150 kW in this embodiment). The upper limit of available power determined by the ramp constraint ( ).

[0224] In the over-limit zone Within that interval, the minimum value of the upper bound of the available power at each moment (i.e., after passing the vehicle's power output capability limit) is taken as the executable power for that interval. : .

[0225] Based on energy demand and available power, service time requirements for: .

[0226] The above calculations yielded and These will be used as constraint parameters in the subsequent joint optimization model.

[0227] Step S4.5: Construct a joint optimization model.

[0228] Define the vehicle set χ, the transformer area set I, and the time period set T. The joint optimization model includes a comprehensive objective function J and constraints. The expression for the comprehensive objective function J is:

[0229] (14)

[0230] in:

[0231] The total travel time for vehicle k is obtained by summing the travel times of each segment from the travel time matrix. .

[0232] The driving energy consumption of vehicle k is obtained by summing the product of the energy consumption per unit distance (0.2 kWh / km in this embodiment) and the distance traveled, i.e. ,in Energy consumption per unit distance This is the l-th segment of the journey.

[0233] The charging and discharging energy loss of vehicle k is calculated as follows: Calculate, where, and These represent the charging efficiency and discharging efficiency of vehicle k, respectively. and Let be the charging power and discharging power of vehicle k at time t, respectively.

[0234] B represents the vehicle load balance, measured by variance: ,in , , This represents the charging or discharging power of vehicle k at time t.

[0235] R represents the vehicle's State of Charge (SOC) margin. , Let vehicle k be at the end of the scheduling cycle. The actual state of charge, Let SOC be the target SOC that vehicle k needs to retain at the end of the scheduling cycle (70% in this embodiment).

[0236] , , , This represents the reference constant for normalization of each term; in this embodiment, it is taken as... =100 min, =50 kWh, =100, =0.2.

[0237] , , , The weight parameters for each item are taken in this embodiment. =1, β=0.5, γ=0.3, μ=0.2.

[0238] Decision variables include:

[0239] : Indicates whether vehicle k is providing service in station i at discrete time t (accessed and not left);

[0240] : Indicates whether vehicle k travels from station i to station j at discrete time t;

[0241] , : The time when vehicle k arrives at and leaves station i (continuous time variable).

[0242] The constraints include:

[0243] Voltage constraint: Planned voltage after considering the role of energy storage vehicles satisfy:

[0244] (15)

[0245] Feasible pairing constraint: Service allocation is only allowed within feasible pairings, i.e., only if (i,j) is within a feasible pairing and the time window requirement is met. If not satisfied, then .

[0246] Service energy constraints: For each over-limit area i and its corresponding over-limit interval The cumulative effective service energy must meet the energy demand:

[0247] (16)

[0248] Service time constraint: The cumulative valid service time must meet the service time requirement.

[0249] (17)

[0250] Energy storage state of charge constraints:

[0251] (18)

[0252] Charge and discharge power constraints:

[0253] (19)

[0254] in, This represents the net output power of vehicle k at time t.

[0255] Reachability constraints:

[0256] (20)

[0257] Where M is a sufficiently large positive number, when Forced time connection.

[0258] Energy reservation constraints:

[0259] (twenty one)

[0260] Step S4.6: Solving and post-processing.

[0261] The above model is solved using a mixed-integer linear programming solver (such as CPLEX or Gurobi). If the solution is infeasible, solution repair and secondary assignment are performed in the following order:

[0262] Power amplitude reduction: Appropriately reduce the executable power. To relax power constraints;

[0263] Replace service area within the set of feasible pairings: Select another service area within the set of feasible pairings to replace the currently infeasible service area;

[0264] Replacement of normal transformer substations: Introducing normal transformer substations as energy transfer nodes;

[0265] Trigger backup distribution area: Activate backup energy storage vehicles or temporary backup distribution areas to participate in dispatching.

[0266] After each repair, the solution is recalculated until a feasible solution is obtained.

[0267] Step S4.7: Generate the day-ahead scheduling scheme.

[0268] Based on the solution results, the day-ahead scheduling plan for each vehicle is generated. For vehicle k, the scheduling plan includes:

[0269] Driving route: The sequence of transformer substations arranged in chronological order and their corresponding driving routes;

[0270] Service sequence: Arrival time of each service counter area Access time departure time ;

[0271] Power dispatch instructions : The planned charge and discharge power sequence at each discrete moment in each service area.

[0272] For vehicle k, its scheduling scheme can be expressed as:

[0273] (twenty two)

[0274] in, Let i be the set of service stations that serve vehicle k sequentially. This is the power scheduling instruction for vehicle k in this area (positive for charging, negative for discharging).

[0275] Output two types of views: a global plan sorted by time period and a local plan for each vehicle, for use in the execution phase of step S5.

[0276] Step S5: Execute the day-ahead dispatching scheme to manage voltage over-limit issues in distribution transformer areas.

[0277] Step S5, according to the day-ahead scheduling plan, controls the energy storage vehicle to travel to the target distribution area and connect to the grid. Through real-time voltage closed-loop correction and rolling replanning, it ensures that the voltage of each distribution area meets the voltage assessment targets within the target time period. Specifically, it includes the following sub-steps:

[0278] Step S5.1: Obtain the power dispatching instruction for the vehicle in the current target area from the day-ahead dispatching plan, and collect the real-time voltage of the target area and the real-time charge status of the vehicle.

[0279] For vehicle k performing a task in transformer area i, obtain the day-ahead dispatch plan for that vehicle from the day-ahead dispatch plan, including: power dispatch instructions. Planned access time and the planned departure time .

[0280] Simultaneously, the real-time low-voltage side voltage of transformer area i at discrete time t is collected in real time through automated charging and discharging piles. The state of charge of vehicle k is collected in real time through the on-board terminal. .

[0281] Step S5.2: Based on the deviation between the real-time voltage and the voltage assessment index, and in combination with the vehicle's power parameters, adjust the power dispatch command in real time to generate the actual power command.

[0282] If the task is charging (the vehicle absorbs power from the transformer area), the target voltage is: Then, based on the deviation between the target voltage and the real-time voltage, the power dispatch command is corrected and limited, and the current actual power command generated is:

[0283] (twenty three)

[0284] If it is a discharge mission (vehicle injecting power into the transformer area), the target voltage is... Then, based on the deviation between the target voltage and the real-time voltage, the power dispatch command is corrected and limited, and the current actual power command generated is:

[0285] (twenty four)

[0286] in, This is the voltage deviation correction factor (0.2 in this embodiment); As a limiting operator, it restricts the current actual power command to the range of vehicle power parameters (including rated charge and discharge power, ramping constraints, and SOC constraints).

[0287] If the current time The vehicle is still in motion, proceeding along the planned route to the target area; if If the task in the current area ends, the process will move on to the next task node.

[0288] Step S5.3: Determine the cumulative effective service energy and time.

[0289] Regarding the current over-limit period Define the effective service energy of vehicle k to the transformer area:

[0290] (25)

[0291] Total effective service time:

[0292] (26)

[0293] Step S5.4: Dynamic update of the day-ahead scheduling scheme for the remaining time period.

[0294] Dynamic updates or rolling replanning are triggered when any of the following conditions occur:

[0295] Voltage deviation exceeds threshold: The deviation between the real-time voltage and the target voltage exceeds the preset voltage deviation threshold. :

[0296] (27)

[0297] Service sequence unreachable: An anomaly occurs in the travel time calculated based on the travel time information, the vehicle access status, or the vehicle status, resulting in the service sequence being unreachable in the day-ahead scheduling plan.

[0298] Insufficient service capacity or time: The accumulated effective service capacity before the planned departure time. Less than the energy demand calculated in step S4.4 or cumulative effective service time Less than the service time requirement .

[0299] Trigger dynamic updates or replanning to calculate the remaining service demand for the current target area:

[0300] Remaining energy requirements: ;

[0301] Remaining service time requirement: .

[0302] Using the current time as the starting time for dynamic updates, Alternative energy demand ,by Alternative service time requirements By combining the vehicle's current location, real-time charge status, information on unserved substations, and updated road network travel time information, step S4 (i.e., regenerating feasible pairing relationships, recalculating executable power, energy demand, and service time demand, and resolving the joint optimization model) is re-executed to obtain the revised day-ahead scheduling scheme.

[0303] Step S5.5: Update the plan and execute iteratively.

[0304] Based on the results of the dynamic update, update the subsequent travel path, service sequence, and power scheduling instructions for vehicle k.

[0305] Repeat steps S5.1 to S5.4 until the real-time voltage of all transformer substations to be treated within the target time period meets the voltage assessment criteria (i.e., ∈[ , ]).

[0306] Example 2

[0307] This invention also provides a distribution transformer area voltage over-limit management system based on an autonomous driving energy storage vehicle. The system includes a memory, a processor, and a computer program or instructions stored in the memory. The processor executes the computer program or instructions to implement the distribution transformer area voltage over-limit management method based on an autonomous driving energy storage vehicle in this invention.

[0308] Although not shown, the system includes a processor that can perform various appropriate operations and processes based on programs and / or data stored in read-only memory (ROM) or loaded from a storage portion into random access memory (RAM). The processor can be a multi-core processor or may contain multiple processors. In some embodiments, the processor may include a general-purpose main processor and one or more specialized coprocessors, such as a central processing unit, graphics processing unit (GPU), neural network processor (NPU), digital signal processor (DSP), etc. Various programs and data required for device operation are also stored in the RAM. The processor, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0309] The processor and memory described above are used together to execute programs / instructions stored in the memory. When the program / instructions are executed by the computer, they can implement the methods, steps, or functions described in the above embodiments.

[0310] The above description only discloses specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or modifications that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for mitigating voltage exceedance in distribution areas based on autonomous driving energy storage vehicles, characterized in that, The method includes: Acquire data on transformer substations, energy storage vehicles, and regional road networks within the target area; the transformer substation data includes at least the voltage assessment indicators, historical voltage data, historical net load data, and day-ahead net load forecast data for each substation; the energy storage vehicle data includes at least the current location, state of charge, and power parameters of each vehicle; the regional road network data includes at least the travel time information of energy storage vehicles between road nodes. Based on the historical voltage data and historical net load data, a voltage prediction model is constructed to map the net load of the distribution area to the voltage of the distribution transformer. According to the voltage prediction model, the day-ahead net load prediction data is converted into day-ahead voltage prediction data, and the day-ahead voltage prediction data is judged to exceed the limit based on the voltage assessment index to obtain the limit exceedance information; the limit exceedance information includes at least the limit exceedance area, the limit exceedance period, and the voltage regulation demand. Based on the over-limit information, energy storage vehicle data, and regional road network data, a joint optimization model is constructed with voltage compliance as a constraint and optimal operating cost as the objective. The joint optimization model is solved to obtain the day-ahead scheduling scheme for each vehicle. The day-ahead scheduling scheme includes at least the travel route, service sequence, and power scheduling instructions for the vehicle in the transformer area. The day-ahead scheduling scheme is implemented to address voltage over-limit issues in distribution transformer areas.

2. The method for managing voltage exceedance in distribution areas based on autonomous driving energy storage vehicles according to claim 1, characterized in that, Constructing the voltage prediction model includes: The historical voltage data and historical net load data are preprocessed to obtain preprocessed sample data; Construct auxiliary features for modeling, the auxiliary features being used to characterize time attributes, environmental attributes and / or operating scenarios; group the sample data according to the auxiliary features, classifying data with the same or similar features into the same group; For each subgroup, the voltage prediction model is determined based on the sample data within that subgroup, and then the hyperparameters of the voltage prediction model are determined; wherein, the distribution transformer voltage is expressed by the following formula: ； in, This represents the low-voltage side voltage of the i-th distribution area at time t, i.e., the distribution transformer voltage. This represents the net load of the i-th transformer area; Indicate auxiliary features; Represents the residual term; Indicates that the hyperparameters are Voltage prediction model; the hyperparameters The estimation is performed by minimizing the objective function, which is expressed as: ； in, This represents the hyperparameter estimates; Indicates sample weights; Represents the loss function; Represents the regularization coefficient; represents the regularization term; T represents the total number of sampling times within the target time period.

3. The method for managing voltage exceedance in distribution areas based on autonomous driving energy storage vehicles according to claim 1, characterized in that, Based on the aforementioned voltage assessment indicators, the day-ahead voltage forecast data is judged to exceed the limit, including: Based on the voltage assessment indicators, set the upper voltage threshold and the lower voltage threshold, and set the upper limit margin and the lower limit margin; For the day-ahead voltage forecast data, the following parameters for each transformer area are statistically analyzed according to the target time period: Duration exceeding the upper limit: The cumulative duration of the day-ahead voltage forecast data ≥ the sum of the upper voltage limit threshold and the upper limit margin; Duration beyond the lower limit: The cumulative duration of the difference between the current day voltage forecast data and the lower voltage limit threshold and the lower limit margin; Percentage of samples exceeding the upper limit: The proportion of times when the current day voltage prediction data is greater than or equal to the sum of the upper limit threshold and the upper limit margin, out of the total number of sampling times; Percentage of samples exceeding the lower limit: The proportion of times when the current day voltage prediction data is ≤ the difference between the lower voltage limit threshold and the lower limit margin out of the total number of sampling times; By combining preset duration thresholds and sample proportion thresholds, the type of each transformer area is determined: If the duration of a transformer area exceeding the upper limit is not less than the corresponding duration threshold, or the proportion of samples exceeding the upper limit is not less than the corresponding sample proportion threshold, it is determined to be an upper limit transformer area; if the duration of a transformer area exceeding the lower limit is not less than the corresponding duration threshold, or the proportion of samples exceeding the lower limit is not less than the corresponding sample proportion threshold, it is determined to be a lower limit transformer area; otherwise, it is determined to be a normal transformer area. For transformer areas that exceed the upper or lower limit, an over-limit indication sequence is generated based on the day-ahead voltage prediction data; the over-limit indication sequence is used to mark whether a voltage over-limit has occurred at each time. Connectivity segment extraction is performed on the over-limit indication sequence to obtain multiple over-limit intervals; Within each over-limit interval, the voltage regulation requirement is calculated, and the over-limit information is output; the voltage regulation requirement represents the amount of voltage regulation required to restore the voltage within that interval to the assessment interval; the assessment interval consists of an upper voltage threshold and a lower voltage threshold.

4. The method for managing voltage exceedance in distribution areas based on autonomous driving energy storage vehicles according to claim 3, characterized in that, The determination of the type of each station area also includes: Hysteresis interval is used to process the judgment boundary; the hysteresis interval is a voltage range defined by a first boundary threshold and a second boundary threshold, wherein the first boundary threshold is the judgment boundary for entering the over-limit state from the normal state, and the second boundary threshold is the judgment boundary for recovering from the over-limit state to the normal state, and the first boundary threshold and the second boundary threshold are different. For the voltage upper limit side, the first boundary threshold is the voltage upper limit threshold, and the second boundary threshold is the voltage upper limit threshold minus a preset hysteresis width. The processing is specifically as follows: When the voltage rises from the assessment range and crosses the first boundary threshold, it is determined that the voltage exceeds the upper limit; It is determined that the voltage has returned to normal only when it drops from the upper limit and crosses the second boundary threshold. For the lower voltage limit side, the first boundary threshold is the lower voltage limit threshold, and the second boundary threshold is the lower voltage limit threshold plus a preset hysteresis width. The specific processing is as follows: When the voltage drops from the assessment range and crosses the first boundary threshold, it is determined that the voltage has exceeded the lower limit. It is determined that the voltage has returned to normal only when it rises from the lower limit and crosses the second boundary threshold.

5. The method for managing voltage exceedance in distribution areas based on autonomous driving energy storage vehicles according to claim 3, characterized in that, After obtaining multiple over-limit intervals, the process also includes: Adjacent out-of-limit intervals are merged; the merging process includes: Calculate the time interval between two adjacent over-limit intervals. If the time interval is less than a preset merging threshold, then merge the two over-limit intervals into one over-limit interval.

6. The method for managing voltage exceedance in distribution areas based on autonomous driving energy storage vehicles according to claim 1, characterized in that, Based on the aforementioned over-limit information, energy storage vehicle data, and regional road network data, a joint optimization model is constructed with voltage compliance as a constraint and optimal operating cost as the objective, including: For each over-limit transformer area, based on its over-limit period and voltage regulation requirements, determine the theoretical power regulation amount required to restore its voltage to meet the voltage assessment indicators, and thereby determine the power regulation direction for that transformer area. Based on the power adjustment direction of each transformer area, cross-limit transformer areas with opposite power adjustment directions are paired with each other, or cross-limit transformer areas that need adjustment are paired with normal transformer areas that do not need adjustment, to generate candidate transformer area mutual assistance pairs. Based on the current location, state of charge, and travel time information of each vehicle, the feasibility of each candidate transformer station mutual assistance pair is screened to obtain feasible pairing relationships; the conditions for feasibility screening include: the vehicle's accessibility and the vehicle's state of charge not being lower than a preset state of charge limit upon arrival. The energy demand is calculated based on the theoretical power adjustment, the executable power is determined based on the vehicle's power parameters and power output capability, and then the service time demand is calculated based on the energy demand and the executable power. The joint optimization model is constructed, which includes a comprehensive objective function J and constraints. The expression for the comprehensive objective function J is: ； in, This represents the total travel time of vehicle k calculated based on travel time information. This represents the driving energy consumption of vehicle k, calculated based on energy consumption per unit mileage and distance traveled. B represents the energy loss during charging and discharging calculated based on the vehicle's charging and discharging efficiency; R represents the vehicle load balance, used to measure the balance of tasks undertaken by each vehicle within the scheduling cycle; and R represents the reserve margin of the vehicle's state of charge. , , , This represents the reference constant for normalization of each term; , , , The weight parameters for each item; The constraints include voltage constraints, feasible pairing constraints, service energy constraints, service time constraints, energy storage state of charge constraints, charge and discharge power constraints, accessibility constraints, and energy reservation constraints.

7. The method for managing voltage exceedance in distribution areas based on autonomous driving energy storage vehicles according to any one of claims 1 to 6, characterized in that, Executing the day-ahead scheduling scheme includes: Control the energy storage vehicle to travel to the target distribution area and connect according to the day-ahead dispatch plan; Real-time data acquisition of the target transformer area's voltage and the vehicle's real-time state of charge; Based on the deviation between the real-time voltage and the voltage assessment index, and in conjunction with the vehicle's power parameters, the power dispatch command is adjusted in real time to generate the actual power command. When a triggering event is detected that requires adjustment of the remaining schedule, the daytime scheduling plan for the remaining period is dynamically updated based on the vehicle's current location, real-time state of charge, power parameters, and remaining over-limit information, and a new scheduling plan is generated. The above real-time adjustment and dynamic update process is executed iteratively until the voltage of each transformer area meets the voltage assessment index within the target time period.

8. The method for managing voltage exceedance in distribution areas based on autonomous driving energy storage vehicles according to claim 7, characterized in that, The real-time adjustment of power scheduling commands includes: Obtain the power dispatching instruction for the vehicle in the current target area from the daytime dispatching scheme; Determine the target voltage based on the type of task the vehicle is currently performing: For charging tasks, the target voltage is the difference between the upper voltage threshold and the preset upper limit margin; for discharging tasks, the target voltage is the sum of the lower voltage threshold and the preset lower limit margin. Based on the deviation between the target voltage and the real-time voltage, the power value at the current moment in the power scheduling command is proportionally corrected to generate the current actual power command. The current actual power command is limited to the range of vehicle power parameters by a limiting operator, resulting in a corrected current actual power command.

9. The method for managing voltage exceedance in distribution areas based on autonomous driving energy storage vehicles according to claim 7, characterized in that, The triggering event includes any of the following: The deviation between the real-time voltage and the target voltage exceeds a preset voltage deviation threshold; the target voltage is determined based on the type of task currently being performed by the vehicle. If the travel time, vehicle access status, or vehicle status calculated based on the travel time information is abnormal, the service sequence in the daytime scheduling scheme will be unreachable. Before the planned departure time, the accumulated effective service energy is less than the energy requirement, or the accumulated effective service time is less than the service time requirement; The effective service energy is the integral of the actual power command during the over-limit period, and the effective service time is the cumulative duration after the vehicle accesses the transformer area.

10. The method for managing voltage exceedance in distribution areas based on autonomous driving energy storage vehicles according to claim 7, characterized in that, The day-ahead scheduling plan for the remaining time period is dynamically updated, including: Calculate the remaining service demand for the current target area, whereby the remaining service demand includes the remaining energy demand and the remaining service time demand. Using the current moment as the starting moment for dynamic updates, and combining the vehicle's current location, real-time state of charge, information on unserved substations, and updated road network travel time information, feasible pairing relationships are regenerated, executable power, energy demand, and service time demand are re-estimated, and the joint optimization model is re-solved to obtain the revised day-ahead scheduling scheme. Based on the revised day-ahead scheduling plan, update the vehicle's subsequent travel routes, service sequence, and power scheduling instructions.

11. A distribution transformer area voltage over-limit management system based on an autonomous driving energy storage vehicle, comprising a memory, a processor, and a computer program or instructions stored in the memory, characterized in that, The processor executes the computer program or instructions to implement the method for managing voltage exceedance in distribution areas based on autonomous driving energy storage vehicles as described in any one of claims 1 to 10.

Images