Intelligent Control Method and System for Charge Stack Power Module Based on Multi-Dimensional State Awareness

Abstract

This invention relates to the field of power management and power dispatch control technology for charging infrastructure, specifically a method and system for intelligent control of charging pile power modules based on multi-dimensional state perception. The method acquires grid-side voltage and power constraints, power module-side output voltage, current, temperature, and operating time, load-side power demand for each charging station, vehicle SOC, and priority; constructs a module state vector, calculates capability assessment values ​​and fatigue assessment values, and classifies modules into high-capacity groups, standard groups, and constrained groups; dynamically groups modules according to group priority under power constraints and generates power allocation for each charging station, determining target power supply to achieve differentiated power supply when necessary; predicts load and constraint changes based on historical data, adjusts assessment parameters and grouping priorities to implement rotation; and issues start / stop and power setpoints to form a closed-loop scheduling system, improving utilization and reducing over-temperature and fault risks. The system includes modules for data acquisition, assessment grouping, grouping allocation, prediction adjustment, and execution.

Description

Technical Field

[0001] This invention relates to the field of power management and power scheduling control technology for charging infrastructure, specifically to a method and system for intelligent control of charging pile power modules based on multi-dimensional state perception. Background Technology

[0002] With the increasing number of new energy vehicles, charging infrastructure has gradually shifted from primarily AC slow charging to DC fast charging, and has continued to evolve in terms of standards and equipment forms. my country established a national standard system for conductive charging systems for electric vehicles as early as around 2001, providing a basic framework for the interconnection and safety requirements of AC and DC chargers. Subsequent standards have been continuously updated to adapt to higher power and more complex off-board power supply equipment application scenarios. In engineering implementation, DC fast charging generally adopts a modular power architecture, centrally configuring multiple charging modules through power cabinets or main units, and then outputting to multiple terminals. The industry has gradually developed charging pile or cluster architectures with "power pooling, terminal sharing, and dynamic allocation" to improve the adaptability and scalability of charging stations. Meanwhile, to meet the needs of supercharging and heavy-duty truck charging, modular parallel operation and higher power level equipment continue to emerge, driving charging piles towards higher power density and more flexible power reconfiguration.

[0003] Existing charging piles or DC fast charging systems still suffer from several pain points in actual operation: On the one hand, the power allocation of some systems remains relatively fixed or follows a single rule, making it difficult to achieve fine-grained, stable, and interpretable scheduling under conditions such as multiple vehicles charging simultaneously, rapidly changing demand, and dynamic adjustments to grid-side power constraints. This easily leads to utilization losses, such as "some charging guns having insufficient power while others are idle." On the other hand, differences in module-level health status and thermal stress objectively exist in engineering, but many systems do not make sufficient use of information such as module temperature and operating time. This may result in some modules operating under high load for extended periods, uneven heating and cooling, and accelerated fatigue accumulation, leading to frequent derating, increased failure rates, and increased maintenance costs. Furthermore, DC fast charging is highly sensitive to grid impacts and station capacity constraints. Without proactive handling of load and constraints, power limiting is often the only option, affecting the charging experience and station revenue. Therefore, a charging pile power module intelligent control method and system based on multi-dimensional state perception is needed to solve these problems. Summary of the Invention

[0004] (a) Technical problem to be solved: In view of the shortcomings of the prior art, the present invention provides a method and system for intelligent control of charging pile power module based on multi-dimensional state perception, which solves the above-mentioned problems.

[0005] (II) Technical Solution: To achieve the above objectives, the present invention provides the following technical solution: a smart control method for power modules of a charging pile based on multi-dimensional state perception, applied to a DC charging pile including multiple power modules and multiple charging guns, the method comprising the following steps: Step S1, acquiring multi-dimensional state information of the grid side, power module side, and load side; the multi-dimensional state information includes at least: grid side state information, including grid voltage and power constraint information; power module side state information, including the first... Output voltage of each power module Output current Module temperature and runtime Load-side status information, including the first The power required by the circuit charging gun The corresponding state of charge of the vehicle The process involves several steps: Step S2, constructing a state vector for each power module based on the multi-dimensional state information, calculating evaluation parameters characterizing the current output capability and fatigue level, and dividing each power module into a high-capacity group, a standard group, and a restricted group according to the evaluation parameters; Step S3, selecting target power modules for dynamic grouping based on the current power demand and charging priority of each charging gun, and combining the high-capacity group, standard group, and restricted group to which each power module belongs, to obtain the target module combination and its power allocation result for each charging gun; Step S4, predicting the charging pile load and power constraints within a preset time window based on historical operating data and current multi-dimensional state information, and adjusting the evaluation parameters and grouping priority according to the prediction results to limit the maximum available power and rotation strategy of each power module in subsequent scheduling cycles; Step S5, issuing start / stop control commands and power setpoints to each power module according to the target module combination and power allocation result to complete the dynamic power management and intelligent control of the charging pile.

[0006] Furthermore, step S2 includes: step S21, regarding the first... Each power module assembles its power module-side state information into a state vector according to preset fields. The state vector At least including output voltage Output current Module temperature and runtime Step S22: Based on the state vector Calculate the evaluation parameters, which at least include those used to characterize the first... Capacity assessment of the current available output capacity of each power module and used to characterize the Fatigue assessment values ​​of accumulated thermal stress during operation of each power module Among them, the ability assessment value With module temperature The fatigue assessment value increases and decreases as the output capacity margin increases. With runtime Increases with module temperature Increase as the value rises; Step S23: Group each power module according to the evaluation parameters, when the capability evaluation value... Meeting the high capability threshold condition and fatigue assessment value When the low fatigue threshold condition is met, the first Each power module is divided into a high-capacity group; when the fatigue assessment value Meet high fatigue threshold conditions or module temperature When the preset temperature threshold is reached, the first One power module was divided into a restricted group; the remaining power modules were divided into a standard group.

[0007] Furthermore, the step S3 of selecting target power modules for dynamic grouping includes: under the premise of satisfying the power constraint information of the grid side, prioritizing the selection of power modules from the high-capacity group, followed by the selection of power modules from the standard group, and only allowing the selection of power modules from the restricted group to participate in grouping when the high-capacity group and the standard group cannot meet the current power demand of each charging gun.

[0008] Furthermore, step S3 also includes determining the target power supply for each charging gun to achieve differentiated power supply for different charging priorities while satisfying the power constraint information. Specifically, this involves generating a target power supply based on the power demand and charging priority of each charging gun, and ensuring that the sum of the target power supply of each charging gun does not exceed the total available power of the charging pile corresponding to the power constraint information. When the total available power of the charging pile is insufficient to meet the power demand of all charging guns, the target power supply of low-priority charging guns is limited or reduced proportionally according to the charging priority until the power constraint information is satisfied.

[0009] Furthermore, the step of limiting or proportionally reducing the target power supply of low-priority charging guns according to charging priority includes: dividing the charging priority of each charging gun into at least two priority levels; when the total available power of the charging pile is insufficient, maintaining the target power supply of the high-priority charging guns at a preset minimum ratio of the corresponding required power, and gradually reducing the target power supply of the low-priority charging guns according to a preset reduction ratio or proportionally allocating it according to the remaining available power, until the sum of the target power supply of each charging gun satisfies the power constraint information.

[0010] Furthermore, the adjustment of the evaluation parameters and grouping priority based on the prediction results in step S4 includes: when the prediction results indicate that the charge pile load will increase or the power constraint will become stricter within the preset time window, increasing the grouping priority of the power modules corresponding to the high-capacity group and decreasing the maximum available power of the power modules corresponding to the restricted group; when the prediction results indicate that the charge pile load will decrease or the power constraint will ease within the preset time window, decreasing the grouping intensity of the power modules corresponding to the high-capacity group and increasing the grouping priority of the power modules corresponding to the standard group, so as to achieve load balancing and fatigue balancing of the power modules.

[0011] Furthermore, the prediction in step S4 includes a strategy recommendation method based on a large-scale pre-trained model, specifically: summarizing the historical operating data and current multi-dimensional state information into a state summary and generating prompt words, which are then input into the large-scale pre-trained model; the large-scale pre-trained model outputs the load change trend or rotation suggestion parameters within the preset time window; the charging pile controller corrects the evaluation parameters or grouping priority according to the rotation suggestion parameters, and the output of the large-scale pre-trained model is not directly issued as the start-stop control command or power setting value.

[0012] Furthermore, the present invention also provides an intelligent control system for a charging pile power module based on multi-dimensional state perception. The system includes: a multi-dimensional state acquisition module for acquiring multi-dimensional state information from the grid side, the power module side, and the load side; the multi-dimensional state information includes at least: grid-side state information, including grid voltage and power constraint information; and power module-side state information, including... Output voltage of each power module Output current Module temperature and runtime Load-side status information, including the first The power required by the circuit charging gun The corresponding state of charge of the vehicle The system includes: a charging priority module; a module evaluation and grouping module, used to construct a state vector for each power module based on the multi-dimensional state information, and calculate evaluation parameters characterizing the current output capability and fatigue level, dividing each power module into a high-capacity group, a standard group, and a restricted group according to the evaluation parameters; a dynamic grouping and power allocation module, used to select target power modules for dynamic grouping based on the current power demand and charging priority of each charging gun, and in combination with the high-capacity group, standard group, and restricted group to which each power module belongs, to obtain the target module combination and its power allocation result for each charging gun; a prediction and strategy adjustment module, used to predict the charging pile load and power constraints within a preset time window based on historical operating data and current multi-dimensional state information, and adjust the evaluation parameters and grouping priority according to the prediction results, so as to limit the maximum available power and rotation strategy of each power module in subsequent scheduling cycles; and a control execution module, used to issue start / stop control commands and power setpoints to each power module according to the target module combination and power allocation result, so as to complete the dynamic power management and intelligent control of the charging pile.

[0013] (III) Beneficial Effects: Compared with the prior art, the present invention provides a charging pile power module intelligent control method and system based on multi-dimensional state perception, which has the following beneficial effects: 1. The charging pile power module intelligent control method and system based on multi-dimensional state perception, through multi-dimensional perception of the power constraints on the grid side, the operating status on the power module side, and the demand and priority on the load side, and constructs a state vector for each power module, further calculates the capability assessment value representing "current available output capability" and the fatigue assessment value representing "operational accumulation and thermal stress", and then divides the power modules into high-capacity group, standard group and restricted group, so that the power scheduling is transformed from the traditional single rule or fixed allocation into interpretable and executable grouped scheduling. Thus, in the scenario of multiple vehicles charging together, rapid changes in demand and dynamic adjustment of grid constraints, it can more accurately identify the modules that can carry high loads and the modules that need to reduce load and rest, reduce the problems of uneven module utilization, uneven heating and cooling and frequent derating, improve the overall operating stability of the charging pile and the consistency of module life, and reduce the risk of failure and operation and maintenance.

[0014] 2. This intelligent control method and system for charging pile power modules based on multi-dimensional state perception achieves dynamic reconfiguration of the power module pool through dynamic grouping rules based on grouping results. When the total power is limited, it determines the target power supply of each charging gun to achieve differentiated power supply for different charging priorities. At the same time, it combines the prediction results of charging pile load and power constraints within a preset time window to adaptively adjust the evaluation parameters and grouping priorities. Thus, when power is tight, it prioritizes high-priority charging tasks and controlsably reduces low-priority tasks, while improving resource release and utilization efficiency when power is ample. The above mechanism transforms scheduling from passive power limiting to proactive planning and adjustment, reduces power allocation fluctuations and reconfiguration frequency, improves charging experience and station revenue, effectively mitigates grid impacts, and ensures the safety, continuity, and stability of the charging process. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the overall process of the intelligent control method for charging pile power modules based on multi-dimensional state perception provided by the present invention.

[0016] Figure 2 This is a schematic diagram of the framework of the intelligent control method for charging pile power modules based on multi-dimensional state perception provided by the present invention.

[0017] Figure 3 This is a schematic diagram of the intelligent control system for charging pile power modules based on multi-dimensional state perception according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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.

[0019] To enable those skilled in the art to better understand the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0020] Please see Figure 1-2 , Figure 1 A schematic diagram of the overall process of the intelligent control method for charging pile power module based on multi-dimensional state perception provided by the present invention; Figure 2This is a schematic diagram of the framework of the intelligent control method for power modules of a charging pile based on multi-dimensional state awareness provided by the present invention. This embodiment provides an intelligent control method for power modules of a charging pile based on multi-dimensional state awareness, applied to a DC charging pile including multiple power modules and multiple charging guns. The method includes the following steps: Step S1, acquiring multi-dimensional state information of the grid side, power module side, and load side; the multi-dimensional state information includes at least: grid side state information, including grid voltage and power constraint information; power module side state information, including the first... Output voltage of each power module Output current Module temperature and runtime Load-side status information, including the first The power required by the circuit charging gun The corresponding state of charge of the vehicle The process involves several steps: Step S2, constructing a state vector for each power module based on the multi-dimensional state information, calculating evaluation parameters characterizing the current output capability and fatigue level, and dividing each power module into a high-capacity group, a standard group, and a restricted group according to the evaluation parameters; Step S3, selecting target power modules for dynamic grouping based on the current power demand and charging priority of each charging gun, and combining the high-capacity group, standard group, and restricted group to which each power module belongs, to obtain the target module combination and its power allocation result for each charging gun; Step S4, predicting the charging pile load and power constraints within a preset time window based on historical operating data and current multi-dimensional state information, and adjusting the evaluation parameters and grouping priority according to the prediction results to limit the maximum available power and rotation strategy of each power module in subsequent scheduling cycles; Step S5, issuing start / stop control commands and power setpoints to each power module according to the target module combination and power allocation result to complete the dynamic power management and intelligent control of the charging pile.

[0021] In this embodiment, the DC charging pile controller, as the unified scheduling unit of the power module pool, first executes step S1 in each scheduling cycle to form three multi-dimensional states that can be used for decision-making: grid constraints, module health, and load demand. The grid side provides power constraint information such as the upper limit of the total available power at the station or the load limit level, while the power module side provides various... Output voltage of each power module Output current Module temperature With runtime The load side provides the first The power required by the circuit charging gun ,vehicle and charging priority; in step S2, the controller uses Construct a state vector and based on and Get the current output power Then, the "available output capacity" and "cumulative thermal stress" are decoupled into evaluation parameters, such as capacity assessment values. Follow The fatigue assessment value increases and decreases as the output capacity margin increases. Follow and The modules are divided into high-capacity, standard, and restricted groups to explicitly indicate which modules are suitable for high loads and which should be prioritized for load reduction and rest. In step S3, the controller first determines the power supply target sequence of each charging gun based on charging priority. Then, under the premise of meeting the power constraints on the grid side, it selects target power modules from the high-capacity group to form target module combinations and supplements each charging gun. If necessary, the standard group supplements the power gap. The restricted group only participates when other modules cannot meet the demand, so as to ensure that critical loads are met while reducing the concentration of thermal risks. In step S4, the controller uses historical operating data and the current status to predict load changes and power constraint changes within a preset time window. When it predicts that the constraints will become stricter or the load will increase, it will reduce the restricted group in advance. The system maximizes available power and prioritizes high-capacity modules in grouping, while implementing a module rotation strategy to avoid prolonged high loads on a single module. When constraints are predicted to ease or loads to decrease, rotation participation is increased to achieve fatigue balancing. Finally, in step S5, the controller maps the target module combination to start / stop control commands and power setpoints for each power module and executes them. After re-acquiring the status in the next scheduling cycle, the system iterates in a closed loop. For example, in a charging pile with multiple charging guns, when the total available power at the station suddenly drops, the system can prioritize power supply to high-priority vehicles and remove high-temperature or high-fatigue modules from the restricted group. At the same time, it can distribute the load to some high-capacity modules through rotation, thereby achieving safe, stable, and interpretable power management in dynamic scenarios of changing grid constraints and multiple vehicles charging simultaneously.

[0022] Furthermore, step S2 includes: step S21, regarding the first... Each power module assembles its power module-side state information into a state vector according to preset fields. The state vector At least including output voltage Output current Module temperature and runtime Step S22: Based on the state vector Calculate the evaluation parameters, which at least include those used to characterize the first... Capacity assessment of the current available output capacity of each power module and used to characterize the Fatigue assessment values ​​of accumulated thermal stress during operation of each power module Among them, the ability assessment value With module temperature The fatigue assessment value increases and decreases as the output capacity margin increases. With runtime Increases with module temperature Increase as the value rises; Step S23: Group each power module according to the evaluation parameters, when the capability evaluation value... Meeting the high capability threshold condition and fatigue assessment value When the low fatigue threshold condition is met, the first Each power module is divided into a high-capacity group; when the fatigue assessment value Meet high fatigue threshold conditions or module temperature When the preset temperature threshold is reached, the first One power module was divided into a restricted group; the remaining power modules were divided into a standard group.

[0023] In this embodiment, the electrical output state, thermal state, and lifetime cumulative state of the module are uniformly mapped into evaluation quantities that can be used for scheduling decisions, thereby transforming complex module differences into executable grouping constraints; specifically, in step S21, the controller targets the first... Each power module will assemble its power module-side state information into a state vector according to preset fields. ,For example ,in and Reflects the module's real-time output capability. Reflects the degree of occupancy of the module's thermal safety boundary. Reflecting the cumulative output power of the module over long-term operation; in step S22, the current output power of the module is first obtained from electrical quantities. and based on the module's rated power Based on temperature Introducing a reduction coefficient Obtain the maximum allowable output power under temperature constraints ,in It can be determined according to segmentation rules, for example, when Time to take ,when Time to take , when Time to take ,in, To preset a safe temperature threshold, A preset temperature threshold is defined; then the output capability margin is defined. and normalize it to To construct capability assessment values ​​in this way ,in , As the preset weighting coefficients, since Increase will decrease And reduce the temperature term, therefore Follow It increases and decreases, and varies with the margin. Increase and increase; at the same time construct fatigue assessment values To characterize the accumulated thermal stress during operation, the operating time is first normalized to... ,in To reference the runtime, a thermal stress term is reconstructed. , when Time to take ,when Time to take And by limiting it to the range of 0 to 1, we can obtain ,in , To preset weighting coefficients, thereby ensuring Follow Increase and grow with Increases with rising; to suppress group jitter, the controller also... Perform recursive smooth update ,in The smoothing coefficient is used; in step S23, grouping is completed according to threshold conditions, for example, when... and The time will be the first Each power module is divided into a high-capacity group, when or They were then divided into restricted groups, and the rest were divided into standard groups. For high capability threshold, To achieve a low fatigue threshold, A high fatigue threshold is required to ensure that the grouping results can directly support the module selection, unloading, and rotation strategies during subsequent dynamic grouping.

[0024] Furthermore, the step S3 of selecting target power modules for dynamic grouping includes: under the premise of satisfying the power constraint information of the grid side, prioritizing the selection of power modules from the high-capacity group, followed by the selection of power modules from the standard group, and only allowing the selection of power modules from the restricted group to participate in grouping when the high-capacity group and the standard group cannot meet the current power demand of each charging gun.

[0025] In this embodiment, the controller first reads the power constraint information from the grid side (e.g., the upper limit of the total available power of the charging pile or the load limit level), and obtains the current power demand and charging priority of each charging gun. Based on this, it determines the set of power supply tasks within the current scheduling cycle. Subsequently, in the module selection phase, the controller executes a hierarchical selection strategy according to the predetermined grouping results: power modules are selected from the high-capacity group first to participate in the grouping, so as to ensure that there is still a good thermal margin and capacity margin when the power output is high, and to reduce the probability of derating and protection triggering. When the number of available modules or the available output of the high-capacity group is insufficient to cover the power demand of all charging guns, power modules are selected from the standard group to make up for the power gap, so as to achieve a smooth distribution of the regular load. For the restricted group, the controller regards it as a risk fallback resource. Only when the high-capacity group and the standard group still cannot meet the current power demand of some charging guns will the modules in the restricted group be allowed to participate in the grouping in a restricted power or restricted duration manner, so as to avoid high-temperature or high-fatigue modules being continuously subjected to high loads, which would lead to accelerated aging or triggering of over-temperature protection. Through the aforementioned dynamic grouping mechanism, the controller can prioritize the allocation of limited module resources to safer and more stable modules in scenarios involving multiple vehicles charging simultaneously, fluctuating demand, and changing power constraints. At the same time, it can isolate and cautiously utilize risky modules, ultimately forming the target module combination and power allocation results for each charging gun, and providing a clear execution basis for subsequent start-stop control and power setpoint issuance.

[0026] Furthermore, step S3 also includes determining the target power supply for each charging gun to achieve differentiated power supply for different charging priorities while satisfying the power constraint information. Specifically, this involves generating a target power supply based on the power demand and charging priority of each charging gun, and ensuring that the sum of the target power supply of each charging gun does not exceed the total available power of the charging pile corresponding to the power constraint information. When the total available power of the charging pile is insufficient to meet the power demand of all charging guns, the target power supply of low-priority charging guns is limited or reduced proportionally according to the charging priority until the power constraint information is satisfied.

[0027] In this embodiment, to ensure that dynamic grouping has a clear power supply target and maintains controllable service trade-offs when station power is limited, step S3 further includes a process for determining the "target power supply". This involves first mapping the power demand and charging priority of each charging gun to a set of target power supplies that can be executed in the current scheduling cycle, and then using this set to drive the selection and allocation of power modules. Specifically, the controller first reads the grid-side power constraint information to obtain the total available power of the charging pile. (This can be a fixed upper limit or an upper limit corresponding to a load limit level), and obtain the required power of each charging gun. Its charging priority, and thus the initial target power supply of each charging gun. When the total available power is sufficient, it can be directly taken. This ensures that user needs are met as much as possible; however, in the event of insufficient total available power, i.e. During this process, the controller provides differentiated power supply based on charging priority: It prioritizes ensuring the target power supply to high-priority charging guns, guaranteeing it is not lower than a preset minimum power or percentage (e.g., not lower than a certain percentage of their required power, or not lower than the minimum power needed to maintain a charging session), and allocates the remaining available power to low-priority charging guns. For low-priority charging guns, the controller uses a "limit" method to set an upper limit on their target power supply, or uses a "proportional reduction" method to uniformly scale their required power, thereby gradually compressing the low-priority load until the constraints are met, ultimately ensuring... The controller transforms the "total power constraint" into "executable targets for each power source," ensuring that high-priority vehicles receive more stable power supply under load or peak constraints, while avoiding frequent fluctuations and user experience degradation caused by disorderly power reduction. The controller then supplies power to these target power sources. As a direct constraint on dynamic grouping and power allocation, the target module combination corresponding to each charging gun outputs according to the target power, thereby achieving differentiated power supply and interpretable scheduling under the power constraint conditions.

[0028] Furthermore, the step of limiting or proportionally reducing the target power supply of low-priority charging guns according to charging priority includes: dividing the charging priority of each charging gun into at least two priority levels; when the total available power of the charging pile is insufficient, maintaining the target power supply of the high-priority charging guns at a preset minimum ratio of the corresponding required power, and gradually reducing the target power supply of the low-priority charging guns according to a preset reduction ratio or proportionally allocating it according to the remaining available power, until the sum of the target power supply of each charging gun satisfies the power constraint information.

[0029] In this embodiment, in order to make the "differentiated power supply when power is insufficient" have directly executable rules and be easy to implement and maintain, the controller refines the "limited by priority or reduced by proportion" into a combination process of tiered and two types of reduction strategies.

[0030] Specifically, the controller first divides the charging priority of each charging gun into at least two priority levels based on the operational strategy or user business attributes, such as a high-priority level and a low-priority level. The high-priority level can be used for emergency vehicles, reserved vehicles, or high-paying users, while the low-priority level can be used for ordinary vehicles or tasks that can be postponed. When the total available power of the charging pile is insufficient, the controller implements a "safety net" rule for the high-priority level, that is, the target power supply of each charging gun in the high-priority level is set to a preset safety net ratio that is no less than its required power, thereby ensuring the availability and power supply stability of critical services. After deducting the safety net power of the high-priority level, the controller allocates the remaining available power to the low-priority level, and can choose one or a combination of two methods: one is a step-by-step reduction method, that is, the target power supply of the low-priority level is reduced in rounds according to a preset reduction ratio, and the total is recalculated after each round of reduction to see if it meets the power constraint, until it does. The second method is the proportional allocation method, which allocates the remaining available power according to the power demand ratio of each charging gun in the low priority level, so that the target power supply of each low priority level can be scaled synchronously under the premise of relative fairness.

[0031] By implementing the aforementioned rules of "priority tiering, ensuring a minimum first and then reducing, with reductions occurring gradually or proportionally," the controller can clearly guarantee the continuity of power supply for high-priority loads when power constraints change abruptly or when multiple vehicles charge simultaneously, leading to resource shortages. At the same time, it can implement interpretable, configurable power compression for low-priority loads that is less likely to cause disorderly fluctuations, thereby providing a stable target power input for subsequent dynamic module grouping and improving the overall controllability of scheduling.

[0032] Furthermore, the adjustment of the evaluation parameters and grouping priority based on the prediction results in step S4 includes: when the prediction results indicate that the charge pile load will increase or the power constraint will become stricter within the preset time window, increasing the grouping priority of the power modules corresponding to the high-capacity group and decreasing the maximum available power of the power modules corresponding to the restricted group; when the prediction results indicate that the charge pile load will decrease or the power constraint will ease within the preset time window, decreasing the grouping intensity of the power modules corresponding to the high-capacity group and increasing the grouping priority of the power modules corresponding to the standard group, so as to achieve load balancing and fatigue balancing of the power modules.

[0033] In this embodiment, step S4 is used to upgrade the charging pile scheduling from "only reacting to the current state" to "preparing in advance for upcoming load and constraint changes", thereby reducing drastic power fluctuations during sudden load limits and achieving a balance between module load and fatigue through planned rotation.

[0034] Specifically, the controller generates predictive inputs based on historical operating data and current multi-dimensional status information, and outputs load trends and power constraint trends within a preset time window. The load trend reflects the increase or decrease in power demand from multiple charging guns, while the power constraint trend reflects changes in the station's load limit level or the tightening or loosening of the upper limit of available total power. When the prediction results indicate that the charging pile load will increase or the power constraint will become stricter within the time window, the controller will, in order to reduce the risk of overheating and derating, advance the participation priority of high-capacity power modules in the grouping, allowing them to enter the target module combination first and undertake higher power output. At the same time, it will impose stricter "maximum available power" limits on restricted power modules or shorten their continuous participation in grouping, avoiding the continued use of high-fatigue or high-temperature modules during the upcoming high-load phase. The module triggers protection; correspondingly, when the prediction result indicates that the load will decrease or the power constraint will ease, the controller will reduce the participation intensity of high-capacity group power modules in the grouping, such as reducing the probability of them being selected into the grouping or reducing their power allocation ratio, and increase the participation priority of standard group power modules in the grouping, so that the standard group can take on more routine output, thereby creating a "rest window" for the high-capacity group to reduce temperature and fatigue indicators and delay aging; through the above-mentioned priority and upper limit adjustment based on prediction, the controller can dynamically switch between "anti-impact mode" and "balanced rotation mode" at different stages, which can ensure stable power supply and safety boundary when the constraint becomes strict, and improve module utilization and fatigue balance when the constraint eases, reducing the inconsistency in lifespan and the increase in failure probability caused by long-term high-load operation of some modules.

[0035] Furthermore, the prediction in step S4 includes a strategy recommendation method based on a large-scale pre-trained model, specifically: summarizing the historical operating data and current multi-dimensional state information into a state summary and generating prompt words, which are then input into the large-scale pre-trained model; the large-scale pre-trained model outputs the load change trend or rotation suggestion parameters within the preset time window; the charging pile controller corrects the evaluation parameters or grouping priority according to the rotation suggestion parameters, and the output of the large-scale pre-trained model is not directly issued as the start-stop control command or power setting value.

[0036] In this embodiment, the prediction may optionally incorporate a strategy recommendation method based on a large-scale pre-trained model. Its core principle is "using a large model to make suggestions and the controller to make decisions." Specifically, the controller first structurally summarizes historical operating data and current multi-dimensional state information to form a state summary for inference. This summary includes, for example, the power demand sequence of each charging gun within a preset time window, the current power constraint level and its recent changes, the grouping results and temperature distribution of each power module, the proportion of restricted groups, and the grouping switching frequency over recent cycles. Subsequently, the controller generates prompts based on a preset template and inputs the state summary into the large-scale pre-trained model, causing it to output load change trend judgments or rotation suggestion parameters within the preset time window. The rotation suggestion parameters may include, for example, the grouping strength to be increased or decreased, the module groups to be prioritized for rotation, and the restriction level for restricted groups. After receiving the output of the large model, the controller does not directly issue it as a start / stop control command or power setpoint. Instead, it uses it as a "soft suggestion" to correct the evaluation parameters or grouping priority. For example, the rotation suggestion is mapped to the priority bias of high-capacity groups, standard groups, and restricted groups, or mapped to the tightening coefficient of the maximum available power of restricted groups. After further superimposing the controller's own hard constraint verification (such as total power not exceeding power constraint information, module temperature not exceeding the threshold, and restricted group calls not violating the restriction conditions), the controller generates the final deterministic scheduling result.

[0037] In this way, the large model is used to absorb historical patterns, abnormal modes and multi-factor coupling relationships, thereby improving the prediction quality of load fluctuations and constraint changes. The controller still retains the final decision-making power and ensures that the output command meets the safety boundary and engineering constraints, avoiding direct control risks caused by the uncertainty of the large model output.

[0038] This invention also provides an intelligent control system for a charging pile power module based on multi-dimensional state perception. The system includes: a multi-dimensional state acquisition module for acquiring multi-dimensional state information from the grid side, the power module side, and the load side; the multi-dimensional state information includes at least: grid-side state information, including grid voltage and power constraint information; and power module-side state information, including... Output voltage of each power module Output current Module temperature and runtime Load-side status information, including the first The power required by the circuit charging gun The corresponding state of charge of the vehicle The system includes: a charging priority module; a module evaluation and grouping module, used to construct a state vector for each power module based on the multi-dimensional state information, and calculate evaluation parameters characterizing the current output capability and fatigue level, dividing each power module into a high-capacity group, a standard group, and a restricted group according to the evaluation parameters; a dynamic grouping and power allocation module, used to select target power modules for dynamic grouping based on the current power demand and charging priority of each charging gun, and in combination with the high-capacity group, standard group, and restricted group to which each power module belongs, to obtain the target module combination and its power allocation result for each charging gun; a prediction and strategy adjustment module, used to predict the charging pile load and power constraints within a preset time window based on historical operating data and current multi-dimensional state information, and adjust the evaluation parameters and grouping priority according to the prediction results, so as to limit the maximum available power and rotation strategy of each power module in subsequent scheduling cycles; and a control execution module, used to issue start / stop control commands and power setpoints to each power module according to the target module combination and power allocation result, so as to complete the dynamic power management and intelligent control of the charging pile.

[0039] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0040] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent control of power modules of a charging pile based on multi-dimensional state perception, applied to a direct-current charging pile comprising a plurality of power modules and a plurality of charging guns, characterized in that, The method includes the following steps: Step S1, acquiring multi-dimensional state information of the grid side, power module side, and load side; the multi-dimensional state information includes at least: grid side state information, including grid voltage and power constraint information; power module side state information, including the first... Output voltage of each power module Output current Module temperature and runtime Load-side status information, including the first The power required by the circuit charging gun The corresponding state of charge of the vehicle And charging priority; Step S2, construct a state vector for each power module according to the multi-dimensional state information, and calculate the evaluation parameters characterizing the current output capability and fatigue level. According to the evaluation parameters, divide each power module into a high-capacity group, a standard group, and a restricted group; Step S3, according to the current power demand and charging priority of each charging gun, and combined with the high-capacity group, standard group, and restricted group to which each power module belongs, select target power modules for dynamic grouping to obtain the target module combination and its power allocation result for each charging gun; Step S4, based on historical operating data and current multi-dimensional state information, predict the charging pile load and power constraints within a preset time window, and adjust the evaluation parameters and grouping priority according to the prediction results to limit the maximum available power and rotation strategy of each power module in the subsequent scheduling cycle; Step S5, issue start-stop control commands and power setpoints to each power module according to the target module combination and power allocation result to complete the dynamic power management and intelligent control of the charging pile; Step S2 includes: Step S21, for the first Each power module assembles its power module-side state information into a state vector according to preset fields. The state vector At least including output voltage Output current Module temperature and runtime Step S22: Based on the state vector Calculate the evaluation parameters, which at least include those used to characterize the first... Capacity assessment of the current available output capacity of each power module and used to characterize the Fatigue assessment values ​​of accumulated thermal stress during operation of each power module Among them, the ability assessment value With module temperature The fatigue assessment value increases and decreases as the output capacity margin increases. With runtime increasing with increasing module temperature increasing with increasing module temperature; step S23, grouping the power modules according to the evaluation parameters, when the capability evaluation value satisfies a high capability threshold condition and the fatigue evaluation value satisfies a low fatigue threshold condition, the first power module is classified into a high capability group; when the fatigue evaluation value satisfies a high fatigue threshold condition or the module temperature reaches a preset temperature threshold, the first power module is classified into a restricted group; the remaining power modules are classified into a standard group.

2. The method for intelligent control of a charging pile power module based on multi-dimensional state perception according to claim 1, characterized in that, The step S3 of selecting the target power module for dynamic grouping comprises: under the premise of meeting the power constraint information of the grid side, the power module is preferentially selected from the high-capability group, then from the standard group, and only when the high-capability group and the standard group cannot meet the current demand power of each charging gun, the power module from the limited group is allowed to participate in the grouping.

3. The method for intelligent control of a multi-dimensional state-aware charging pile power module according to claim 1, characterized in that, The step S3 further comprises determining the target power supply of each charging gun to realize differentiated power supply for different charging priorities under the premise of meeting the power constraint information, specifically: generating the target power supply based on the demand power and charging priority of each charging gun, and making the sum of the target power supply of each charging gun not exceed the total available power of the charging pile corresponding to the power constraint information; when the total available power of the charging pile is insufficient to meet the demand power of all charging guns, the target power supply of low-priority charging guns is adjusted by limit or proportion according to the charging priority, until the power constraint information is met.

4. The method for intelligent control of a charging pile power module based on multi-dimensional state perception according to claim 3, characterized in that, The step of adjusting the target power supply of low-priority charging guns by limit or proportion according to the charging priority comprises: dividing the charging priority of each charging gun into at least two priority levels; when the total available power of the charging pile is insufficient, the target power supply of high-priority charging guns is kept not lower than a preset bottom line proportion of the corresponding demand power, and the target power supply of low-priority charging guns is gradually reduced by a preset adjustment proportion or proportionally distributed according to the remaining available power, until the sum of the target power supply of each charging gun meets the power constraint information.

5. The method for intelligent control of a multi-dimensional state-aware charging pile power module according to claim 1, characterized in that, The step S4 of adjusting the evaluation parameters and grouping priority according to the prediction result comprises: when the prediction result represents that the charging pile load will rise or the power constraint will become more stringent within the preset time window, the participation grouping priority of the power module corresponding to the high-capability group is increased and the maximum available power of the power module corresponding to the limited group is reduced; when the prediction result represents that the charging pile load will decrease or the power constraint will become less stringent within the preset time window, the participation grouping intensity of the power module corresponding to the high-capability group is reduced and the participation grouping priority of the power module corresponding to the standard group is increased, to realize load balancing and fatigue balancing of the power module.

6. The method for intelligent control of a multi-dimensional state-aware charging pile power module according to claim 1, characterized in that, The prediction in the step S4 comprises a strategy recommendation mode based on a large-scale pre-training model, specifically: the historical running data and the current multi-dimensional state information are summarized as a state summary and a prompt word is generated and input to the large-scale pre-training model, and the large-scale pre-training model outputs the load change trend or rotation suggestion parameter within the preset time window; The charging pile controller corrects the evaluation parameters or grouping priority according to the rotation suggestion parameter, and the output of the large-scale pre-training model is not directly issued as the start-stop control instruction or power setting value.

7. A charging stack power module intelligent control system based on multi-dimensional state perception, characterized in that, The system is used to realize the intelligent control method of the charging pile power module based on multi-dimensional state perception according to any one of claims 1 to 6.

Images