CN-121984059-A - Flywheel lithium battery hybrid energy storage capacity configuration method and system

CN121984059ACN 121984059 ACN121984059 ACN 121984059ACN-121984059-A

Abstract

The invention discloses a flywheel lithium battery hybrid energy storage capacity configuration method and system, which particularly relate to the field of flywheel lithium batteries, and comprise the steps of S1, adopting multi-sensor fusion to collect multidimensional operation parameters, carrying out data preprocessing through a Kalman filtering algorithm, S3, constructing a reinforcement learning collaborative model, obtaining equipment attenuation efficiency, optimizing and outputting the optimal power distribution proportion of a flywheel and a lithium battery through model training, S5, constructing an attention mechanism model to predict future comprehensive power fluctuation, optimizing prediction precision, S7, obtaining real-time actual capacity, setting dual triggering conditions, updating the reinforcement learning model and dynamically adjusting capacity, and realizing accurate perception and prediction of new energy fluctuation and load demand through a multi-sensor fusion collection technology, a wavelet packet decomposition fluctuation extraction technology and an attention mechanism reinforced LSTM prediction technology, and solving the problem of low configuration precision caused by the fact that the prior art depends on an empirical formula and a static model.

Inventors

Dou Cai
CHEN LIANG
GAO LIN
WANG WEISHEN
WANG GANG
LI TIANHUI

Assignees

辽宁大唐国际新能源有限公司

Dates

Publication Date: 20260505
Application Date: 20251224

Claims (10)

1. The flywheel lithium battery hybrid energy storage capacity configuration method is characterized by comprising the following steps: S1, acquiring multidimensional operation parameters by adopting multi-sensor fusion, and preprocessing data by using a Kalman filtering algorithm; S2, decomposing by adopting wavelet packets, extracting energy entropy of each fluctuation component, determining stabilizing requirements and distributing the fluctuation components according to frequency characteristics; s3, constructing a reinforcement learning cooperative model, acquiring equipment attenuation efficiency, and optimizing and outputting the optimal power distribution ratio of the flywheel and the lithium battery through model training; S4, building a digital twin body of the hybrid energy storage system, building a full life cycle cost model, setting constraint conditions and carrying out cost simulation; s5, constructing an attention mechanism model to predict future comprehensive power fluctuation, and simultaneously optimizing prediction accuracy; s6, pre-distributing initial capacity based on the fluctuation component characteristics, simultaneously performing cost verification, and determining final initial capacity by combining constraint conditions; S7, acquiring real-time actual capacity, setting dual triggering conditions, updating the reinforcement learning model and dynamically adjusting the capacity; s8, building a virtual-real combination verification platform, synchronously operating data, calculating the comprehensive efficiency and the full life cycle cost of the system, and verifying the validity of the configuration scheme.
2. The method of claim 1, wherein the multi-dimensional operating parameters include new energy side parameters, load side parameters, flywheel device parameters, lithium battery device parameters and system constraint parameters, and the new energy side parameters include photovoltaic output power instantaneous values Instantaneous value of wind power output power The load side parameter specifically includes real-time power demand Power fluctuation allowable threshold The flywheel equipment parameters specifically comprise maximum charge and discharge power Charge and discharge efficiency Response time Total number of cycle life Number of used cycles The lithium battery equipment parameters specifically comprise maximum charge and discharge power Cost of construction Cost per annual capacity loss Annual maintenance cost per unit capacity Attenuation coefficient Continuous run time Maximum charge-discharge power of lithium battery equipment Charge and discharge efficiency Response time Total number of cycle life Number of used cycles Cost per unit capacity of construction Cost per annual capacity loss Annual maintenance cost per unit capacity Attenuation coefficient Continuous run time The system constraint parameters comprise the full life cycle period T, the discount rate r and the capacity attenuation threshold value of the energy storage system Prediction accuracy threshold Threshold of system comprehensive efficiency Full lifecycle cost threshold 。
3. The method for configuring the hybrid energy storage capacity of the flywheel lithium battery as claimed in claim 2, wherein the wavelet packet decomposition is performed by using db4 basis functions Performing 3-layer decomposition to separate high-frequency instantaneous fluctuation and low-frequency continuous fluctuation, wherein the fluctuation component energy entropy comprises photovoltaic fluctuation component energy entropy Wind power fluctuation component energy entropy Load fluctuation component energy entropy Wherein Representing the energy duty cycle of the ith component of the photovoltaic fluctuation, Representing the energy duty cycle of the ith component of the wind power fluctuation, Representing the energy duty ratio of the ith component of the load fluctuation, wherein the stabilizing requirement is realized by integrating the power fluctuation value And power exceeding part And determining that the fluctuation component distribution rule distributes high-frequency fluctuation components to flywheel responses and low-frequency fluctuation components to lithium battery responses.
4. The method for configuring the hybrid energy storage capacity of the flywheel lithium battery as claimed in claim 2, wherein the reinforcement learning collaborative model comprises a state space definition: The action space is defined as the output power ratio of flywheel Output power ratio of lithium battery The bonus function defines: wherein In order to be of a synergistic efficiency, The damping efficiency of the equipment comprises the actual charge and discharge efficiency after the flywheel is damped And actual charge-discharge efficiency after lithium battery decay The model training adopts a DQN algorithm, the iteration times are more than or equal to 5000 times, the convergence error is less than or equal to 1%, and the optimal power distribution proportion After output, determining the actual output power of the flywheel Actual output power of lithium battery And satisfies the power constraint.
5. The method for configuring the hybrid energy storage capacity of the flywheel lithium battery as set forth in claim 4, wherein the digital twin technology maps physical properties, operating states and environmental impact factors of the device in real time by building a digital twin body of the flywheel-lithium battery hybrid energy storage system, synchronizes attenuation data and operation loss data of the device, and the full life cycle cost model Wherein Represents the energy storage capacity of the flywheel, Represents the energy storage capacity of the lithium battery, Indicating the predicted damping efficiency of the flywheel at different times, The predicted decay efficiency of the lithium battery at different moments is shown, Representing full lifecycle costs, the constraints include power constraints Capacity constraint Stabilization constraint Life-constrained full life cycle content decay 。
6. The method of claim 2, wherein the attention mechanism model divides the historical data into a training set and a test set according to a 7:3 ratio by the enhanced LSTM model, the input sequence length is set to 24, and the attention mechanism layer passes through the formula Giving feature weights to different moments of the input sequence, and predicting future time window by using model Integrated power fluctuations within By mean absolute error Evaluating accuracy, wherein , Representing the input feature at time t in the attention mechanism, Represents an attention mechanism training parameter, b represents an attention mechanism training parameter, Representing feature scores at time t in the attention mechanism, Representing predicted integrated power fluctuations within a future time window, The optimized prediction precision is adjusted by adopting an adaptive learning rate, the initial learning rate is 0.001, and the attenuation is 10% for every 100 iterations until 。
7. The method for configuring the hybrid energy storage capacity of a flywheel lithium battery as defined in claim 4, wherein said pre-allocated capacity comprises an initial pre-allocated capacity of the flywheel Initial pre-allocation capacity of lithium battery The verification cost process calculates the pre-allocation capacity corresponding to the digital twin cost model If (if) According to the adjustment coefficient The pre-allocation capacity is adjusted, and the final initial capacity is determined by substituting the adjusted pre-allocation capacity into the constraint condition in S4, and re-optimizing the power allocation proportion through a reinforcement learning model if any constraint is not satisfied Repeating S1-S2 until all constraints are satisfied, and outputting final initial capacity 。
8. The method for configuring the hybrid energy storage capacity of the flywheel lithium battery as set forth in claim 7, wherein the method for obtaining the real-time actual capacity is characterized by collecting real-time operation data of the device every 5 minutes through a dynamic adjustment mechanism driven by edge calculation to obtain the real-time actual capacity of the flywheel And real-time actual capacity of lithium battery The triggering condition is Or (b) And (2) and The capacity adjustment is re-optimized by updating the reinforcement learning model state variable Adjusting to obtain the dynamic adjusted capacity of the lithium battery Wherein Representing the efficiency of the flywheel decay at the initial configuration, Representing the lithium battery decay efficiency at initial configuration, Represents the real-time decay efficiency of the lithium battery, Representing the real-time decay efficiency of the lithium battery.
9. The flywheel lithium battery hybrid energy storage capacity configuration method is characterized in that the virtual-real combination verification platform comprises a physical end and a virtual end, the physical end builds a small hybrid energy storage experimental platform, a simulation new energy power generation system and a load system are accessed, the virtual end calls digital twin body synchronous physical end operation data, and the verification configuration scheme is effective in the following manner: and the performance error of the physical end and the virtual end is less than or equal to 3%, otherwise, the readjustment is returned.
10. A flywheel lithium battery hybrid energy storage capacity configuration system for implementing a flywheel lithium battery hybrid energy storage capacity configuration method according to any of the preceding claims 1-9, comprising: the initial module adopts multi-sensor fusion to collect multi-dimensional operation parameters and performs data preprocessing through a Kalman filtering algorithm; the extraction and distribution module is used for extracting energy entropy of each fluctuation component by adopting wavelet packet decomposition, determining stabilizing requirements and distributing the fluctuation components according to frequency characteristics; The reinforcement learning module is used for constructing a reinforcement learning cooperative model, acquiring equipment attenuation efficiency, and optimizing and outputting the optimal power distribution ratio of the flywheel and the lithium battery through model training; the generation cost constraint module is used for constructing a digital twin body of the hybrid energy storage system, constructing a full life cycle cost model, setting constraint conditions and carrying out cost simulation; The attention prediction module is used for constructing an attention mechanism model to predict future comprehensive power fluctuation and optimize prediction precision; the pre-allocation verification initial allocation module is used for pre-allocating initial capacity based on the fluctuation component characteristics and simultaneously performing cost verification, and determining final initial capacity by combining constraint conditions; The edge calculation module is used for acquiring real-time actual capacity, setting dual triggering conditions, updating the reinforcement learning model and dynamically adjusting the capacity; And the scheme verification module is used for building a virtual-real combination verification platform, synchronously running data, calculating the comprehensive efficiency and the full life cycle cost of the system and verifying the validity of the configuration scheme.

Description

Flywheel lithium battery hybrid energy storage capacity configuration method and system Technical Field The invention relates to the technical field of flywheel lithium batteries, in particular to a flywheel lithium battery hybrid energy storage capacity configuration method and system. Background With the large-scale application of new energy power generation technologies such as photovoltaic, wind power and the like, the intermittence and fluctuation of output power of the system have remarkable influence on the stable operation of a power grid, and the flywheel energy storage and lithium battery hybrid energy storage system is widely applied to the scenes such as power fluctuation stabilization, peak clipping and valley filling of the power grid, emergency power supply and the like of the new energy power generation side by virtue of the characteristics of high flywheel response speed, long cycle life, no obvious limitation on charge and discharge times, high energy density of the lithium battery, large energy storage capacity and stable charge and discharge characteristics, and becomes an important technical means for solving the problem of the stability of new energy grid connection. The existing flywheel lithium battery hybrid energy storage capacity configuration relieves the performance short plates of a single energy storage system to a certain extent by combining the characteristics of two energy storage devices, can realize complementary supply of power and energy, and improves the stabilizing capability of the energy storage system on new energy fluctuation. The intelligent energy storage system has the obvious defects that firstly, capacity configuration mostly depends on a simplified empirical formula or static model, dynamic changes of new energy output power and load demands and attenuation characteristics in the operation process of energy storage equipment are not fully considered, an accurate real-time data sensing and dynamic adaptation technology is lacked, so that a configuration scheme is not matched with actual operation demands, secondly, an accurate quantitative modeling and intelligent allocation strategy for the collaborative operation efficiency of a flywheel and a lithium battery is lacked, the capacities of the flywheel and the lithium battery are simply overlapped, collaborative loss caused by response characteristic differences among the equipment is ignored, the overall operation efficiency of the system is reduced, thirdly, the total life cycle cost is not included in core constraint of the capacity configuration, and a cost dynamic prediction means based on digital twin is lacked, only initial construction cost is concerned, so that maintenance cost and loss cost are overhigh in the long-term operation process, economy is poor, and fourthly, an dynamic adjustment mechanism and edge calculation support are lacked, real-time optimization cannot be carried out according to equipment states and load changes after the configuration scheme is determined, and adaptability and flexibility are insufficient. Aiming at the problems of low configuration accuracy, poor cooperative efficiency, insufficient economy and weak adaptability in the prior art, the invention provides a flywheel lithium battery hybrid energy storage capacity configuration method and system, and the accurate, efficient and economic configuration of the capacity of a hybrid energy storage system is realized through multi-dimensional parameter accurate acquisition, fluctuation characteristic intelligent extraction, cooperative strategy dynamic optimization, full life cycle cost simulation, load prediction enhancement, capacity intelligent configuration, real-time dynamic adjustment and virtual-real combination verification. Disclosure of Invention In order to overcome the above-mentioned drawbacks of the prior art, embodiments of the present invention provide a flywheel lithium battery hybrid energy storage capacity configuration method and system, which solve the problems set forth in the above-mentioned background art through the following schemes. In order to achieve the purpose, the invention provides the following technical scheme that the flywheel lithium battery hybrid energy storage capacity configuration method comprises the following steps: S1, acquiring multidimensional operation parameters by adopting multi-sensor fusion, and preprocessing data by using a Kalman filtering algorithm; S2, decomposing by adopting wavelet packets, extracting energy entropy of each fluctuation component, determining stabilizing requirements and distributing the fluctuation components according to frequency characteristics; s3, constructing a reinforcement learning cooperative model, acquiring equipment attenuation efficiency, and optimizing and outputting the optimal power distribution ratio of the flywheel and the lithium battery through model training; S4, building a digital twin body of the hybrid energy storage system,