CN-121261356-B - Zero-loss energy routing method and system for base station energy storage system

CN121261356BCN 121261356 BCN121261356 BCN 121261356BCN-121261356-B

Abstract

The invention relates to the technical field of base station energy management, and discloses a zero-loss energy routing method and system of a base station energy storage system. The method comprises the steps of creating a virtual model of a base station energy storage system, collecting energy parameter instant data by using monitoring equipment, updating the energy parameter instant data to the model in real time, simulating energy flow in the model to predict a demand change trend and possible loss, integrating an energy utilization rate prediction index and a load balance evaluation value to form a characteristic combination when a route control scheme is designed, detecting the risk of inconsistent logic of a route rule by using an analysis model, comparing the detection value with a standard value to judge whether adjustment is carried out, calling a high-performance switch array by an energy router to control the electric energy flow direction based on the adjusted scheme, automatically executing energy scheduling action without boosting, reducing voltage, inverting and rectifying operation, and recording operation response for continuously optimizing the virtual model. The system comprises a model creation module, a state simulation module, a strategy design module, an operation execution module and a model optimization module.

Inventors

WU HUANBIN
YANG ZHIPENG
CHEN TONGWEI
FU ZESEN

Assignees

广东云山能源科技有限公司

Dates

Publication Date: 20260512
Application Date: 20251202

Claims (9)

1. A zero-loss energy routing method of a base station energy storage system is characterized by comprising the steps of creating a virtual model of the base station energy storage system, collecting instant data of energy parameters by using monitoring equipment, and updating the instant data into the virtual model in real time; The method comprises the steps of designing a route control scheme, integrating the interrelation of an energy utilization rate prediction index and a load balance evaluation value to form a characteristic combination, adopting an analysis model to detect inconsistent risks in a route rule logic according to a preset operation sequence, comparing the rule logic inconsistent risk value with a standard value, judging whether adjustment is needed or not, calling a high-performance switch array to control the electric energy flow direction through an energy router based on the adjusted control scheme, automatically executing energy scheduling action, not performing boosting, reducing, inverting and rectifying operations, and recording operation response for continuously optimizing a virtual model; The method comprises the steps of dividing an energy storage system into a plurality of monitoring areas, establishing node connection based on an energy flow path or a heat conduction relation, associating a plurality of instant data with each node to form a node characteristic set, constructing an initial network structure, representing a system state of each time point as a network state sequence, combining node characteristic change and network structure stability to form time sequence input data, extracting node time-space characteristics by using a time sequence network model, calculating the network state by using the time sequence network model to obtain node embedding representation, estimating future energy loss variables of each node by using the node embedding representation input prediction module, carrying out abnormal evaluation on estimated results by combining history distribution, and marking the estimated results as potential energy loss if the estimated results deviate from a normal range; The rule logic inconsistent risk value is compared with a standard value, and judging whether adjustment is needed or not comprises the steps of comparing the acquired rule logic inconsistent risk value with the standard value, generating an alarm signal and optimizing a route control scheme if the risk value is greater than or equal to the standard value, and not generating the alarm signal if the risk value is smaller than the standard value, so that additional adjustment is not needed.
2. The zero-loss energy routing method of the base station energy storage system according to claim 1 is characterized in that the base station energy storage system adopts an energy pool structural design, integrates power supplies with different voltages and different performances and storage battery access of different types, correspondingly sets multi-source access nodes and hybrid energy storage nodes in a virtual model, increases power supply voltage level, performance parameters, access capacity and battery type parameters in a node characteristic set, configures an energy router, realizes distribution of electric energy according to load demands by the multi-power supply through the energy router, automatically adjusts an energy transmission path by adopting a dynamic routing mechanism according to real-time output characteristics of the different power supplies, charge and discharge demands of the battery and load power demands, realizes intelligent distribution of the electric energy between the multi-source and the multi-load, integrates an anti-backflow module and a low-loss circuit structure in the energy transmission path, supports input anti-backflow of all the access energy storage devices, can effectively prevent unreasonable loss of the electric energy, carries out simulation by setting backflow detection nodes and loss coefficient parameters in the virtual model, optimizes state data of the anti-backflow module and resistance and reactance parameter data of the low-loss circuit structure, and is used for optimizing a routing strategy to predict energy loss.
3. The method of claim 1, wherein the instant data includes power level of the energy storage unit, power demand, temperature parameters and environmental impact factors, and real-time power consumption data of the energy router, switch array status of the energy router, operation status of the anti-reverse flow module, and each load.
4. The zero-loss energy routing method of the base station energy storage system according to claim 1 is characterized in that the energy utilization rate prediction index is obtained by building an energy utilization causal graph, determining variable nodes related to energy utilization rate, including energy storage unit temperature, load level, electric quantity state and scheduling quantity, defining causal dependency relationship among variables to form a directed acyclic graph, setting the energy utilization rate prediction index as a target node, obtaining historical energy data for training a probability network, collecting variable observed values in real time in operation as an input node to input the probability network, triggering a conditional propagation mechanism in the probability network by using the input node, updating posterior probabilities of all related nodes by the network, and obtaining probability distribution results of the energy utilization rate prediction index nodes as current energy utilization rate prediction indexes.
5. The zero-loss energy routing method of the base station energy storage system according to claim 4 is characterized in that a high-precision load sensor is deployed on the surface layer of the energy storage system, power data of loads are collected at fixed intervals, each data sequence represents a load change track in a fixed window to form a load time sequence sample, an encoder is used for carrying out unsupervised feature extraction on the load time sequence, the encoder part automatically extracts low-dimensional feature vectors representing load change modes in each sequence, the decoder part restores the feature vectors to an original load curve, after training is finished, the encoder compresses new collection sequences to feature representations, a grouping algorithm is used for grouping the feature vectors, each group represents a typical load mode, load balancing labels are distributed for each type of mode after grouping, in operation, the system continuously collects current load time sequences, feature compression is carried out by using the trained encoder, the feature vectors are matched with a grouping center, the current sequence belongs to the mode is judged, and the current load balancing evaluation value is output according to the type.
6. The method of claim 5, wherein the energy utilization prediction index and the load balancing evaluation value are converted into feature combinations, the feature combinations are used as inputs of an analysis model, the analysis model is based on a currently set operation sequence, each set of feature combination prediction rule logic inconsistency risk value labels is based on a training target, a prediction error sum of all rule logic inconsistency risk value labels is minimized, the analysis model is trained until the error sum converges, and rule logic inconsistency risk values are determined according to model outputs, wherein the analysis model is a classification tree model.
7. The zero-loss energy routing method of the base station energy storage system according to claim 1, wherein continuously optimizing the virtual model based on the recorded operation responses comprises creating a virtual model simulation energy flow process in an initial stage, simultaneously creating a data recording structure, associating each operation response with a corresponding relation of model output, constructing a plurality of task sets by using historical operation responses, wherein each task represents a specific energy scene and comprises a training subset and a verification subset, performing gradual fine tuning on the virtual model in each task, iteratively optimizing model parameters through the plurality of tasks, constructing an operation response acquired after each operation into a new task in actual operation, inputting the new task into a trained meta-learning framework for iteration, performing local fine tuning on the current virtual model, reserving the fine tuning model as a current scene sub-model, periodically verifying sub-model prediction performance by the system, and if errors reduce or meet a threshold value, extracting fine tuning parameter changes by the system, updating a main model parameter set or adding the main model library.
8. The zero-loss energy routing method of the base station energy storage system according to claim 1 is characterized by further comprising the steps that the edge terminal uploads local energy data to the cloud server, the local energy data comprise node states and preliminary scene types, the cloud server performs grouping analysis on the local energy data and the preliminary scene types to obtain node state grouping centers and scene type preliminary grouping centers, performs space structure optimization on the node state grouping centers based on physical layout of the energy storage system to obtain optimized node state grouping centers, fuses the scene type preliminary grouping centers and the optimized node state grouping centers to obtain energy scene fusion characteristics, determines energy scene type labels based on the energy scene fusion characteristics, matches preset routing strategies from a strategy base based on the energy scene type labels, and sends the routing strategies to the corresponding edge terminal.
9. A zero-loss energy routing system of a base station energy storage system, for implementing the zero-loss energy routing method of the base station energy storage system according to any one of claims 1 to 8, comprising: The model creation module is used for building a virtual model of the base station energy storage system, acquiring instant data of energy parameters through the monitoring equipment and updating the virtual model in real time; The state simulation module is used for simulating the energy flow state in the virtual model and predicting the energy demand change trend and the potential energy loss; The strategy design module is used for designing a route control scheme, integrating the energy utilization rate prediction index and the load balance evaluation value to form a characteristic combination, adopting an analysis model to detect rule logic inconsistent risks according to an operation sequence, and comparing the rule logic inconsistent risk value with a standard value to judge adjustment requirements; the operation execution module comprises an energy router, a high-performance switch array and an anti-reflux module, and is used for calling the high-performance switch array through the energy router to control the electric energy flow direction based on the adjusted control scheme, executing energy scheduling actions and recording operation responses; and the model optimization module is used for continuously optimizing the virtual model based on the operation response.

Description

Zero-loss energy routing method and system for base station energy storage system Technical Field The invention relates to the technical field of base station energy management, in particular to a zero-loss energy routing method and system of a base station energy storage system, which are particularly suitable for high-efficiency energy distribution under multi-power and multi-load scenes. Background With the rapid development of mobile communication technology, the number of base stations serving as key nodes of a communication network is continuously increasing, and higher requirements are placed on the stability, compatibility and zero-loss transmission of energy supply. The base station energy storage system is used as a core component for guaranteeing uninterrupted operation of the base station, power supplies with different voltages and different performances are required to be simultaneously adapted to be connected, precise electric energy distribution of multiple loads is required to be realized, and the current system has obvious defects in the aspects of multi-source compatibility, zero-loss scheduling and backflow prevention protection. The conventional energy routing method of the base station energy storage system has three major core problems that firstly, the multi-source access compatibility is poor, the power supplies with different voltage classes and different performances cannot be matched for power supply in a cooperative manner, the energy loss is obvious due to the fact that the power supplies are required to be subjected to conversion operations such as boosting, reducing and inverting, a unified core control unit is lacking, the power distribution between multiple power supplies and multiple loads depends on a fixed strategy, a transmission path cannot be adjusted according to the real-time output and load dynamic requirements of the power supplies, the phenomenon of insufficient load power supply or power supply power discarding is easy to occur, thirdly, the backflow prevention protection is incomplete, only the backflow prevention measures are set for part of the power supplies, the risk of unreasonable loss of the electric energy exists, the running states of the backflow prevention module and the core control unit are not brought into a monitoring range, the data collection is incomplete, and an accurate system running model is difficult to construct. In the prior art, the energy scheduling process also faces the contradiction between routing logic consistency and zero loss, namely, although the traditional method tries to improve efficiency by optimizing energy utilization rate or load balancing, the traditional method does not combine zero loss demand of 'no conversion operation' to design routing rules, lacks logic risk detection of a core control unit, is easy to generate rule conflict, and further aggravates energy loss or equipment fault risk. In addition, the manual intervention is low in efficiency, simulation verification links are omitted, so that the landing risk of a new routing strategy is high, and the requirement of a base station on efficient energy utilization is difficult to meet. Manual intervention in the traditional energy scheduling process is not only low in efficiency, but also unreasonable in scheduling due to human judgment errors, and further energy loss is aggravated. The lack of effective simulation verification links makes the new routing strategy unable to fully check the feasibility and rationality before practical application, and increases the risk of system operation. Due to the existence of the problems, the energy utilization efficiency of the current base station energy storage system is generally low, the energy waste phenomenon is serious, the operation cost is increased, and the current energy conservation and emission reduction development concept is contrary. Therefore, a method for realizing zero-loss energy routing is needed to improve the operation efficiency and stability of the base station energy storage system. Disclosure of Invention The invention aims to provide a zero-loss energy routing method of a base station energy storage system, so as to solve the problems in the background technology. In order to achieve the above object, the present invention provides a zero-loss energy routing method of a base station energy storage system, the method comprising: creating a virtual model of the base station energy storage system, collecting instant data of energy parameters by using monitoring equipment, and updating the instant data into the virtual model in real time; The method comprises the steps of designing a route control scheme, integrating correlations between an energy utilization rate prediction index and a load balance evaluation value to form a characteristic combination, adopting an analysis model to detect inconsistent risks in route rule logic according to a preset operation sequence, comparing a detection value with