Lithium ion battery management system and method for electric motorcycle

Abstract

The invention discloses an electric motorcycle lithium ion battery management system and method, and relates to the technical field of electric motorcycle battery management, the system comprises the following components: a data acquisition module, a digital twin model construction module, a simulation analysis module and a strategy making module; by integrating the data acquisition module, the digital twinborn model construction module, the simulation analysis module, the strategy making module and the like, the system can acquire battery operation data in real time, a digital model consistent with the state of an entity battery is constructed in a virtual space, and the system can simulate the performance change of the battery under different working conditions and environments by utilizing the model, so that the performance of the battery is improved. Therefore, the service life attenuation condition of the battery is accurately predicted, accurate battery maintenance and replacement strategy suggestions are provided for users, and the accuracy and efficiency of battery management are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric motorcycle battery management, and specifically to an electric motorcycle lithium-ion battery management system and method. Background Art

[0002] With the rapid development of the electric motorcycle market, as the core power source, the performance state of lithium-ion batteries is directly related to the endurance and service life of electric motorcycles. However, the performance of lithium-ion batteries is affected by various factors, including working conditions, ambient temperature, and usage frequency. Changes in these factors may lead to battery performance degradation, thereby affecting the overall performance of electric motorcycles.

[0003] In traditional technologies, the management of lithium-ion batteries for electric motorcycles mainly relies on simple voltage and current monitoring and regular manual inspections. This management method has many deficiencies. On the one hand, it cannot obtain real-time and accurate performance data of the battery under various working conditions and environments, resulting in inaccurate prediction of battery life attenuation. On the other hand, due to the lack of a scientific basis for formulating maintenance strategies, battery maintenance and replacement often can only be carried out according to experience. This not only increases maintenance costs but also may affect the service life and safety of the battery due to untimely or excessive maintenance. In addition, the interaction method between users and the battery management system in traditional technologies is single, lacking an intuitive and convenient information display and warning mechanism, making it difficult for users to timely understand the battery status and make corresponding treatments.

[0004] In view of the many deficiencies of traditional technologies in the management of lithium-ion batteries for electric motorcycles, therefore, it is particularly important to develop an electric motorcycle lithium-ion battery management system and method. Summary of the Invention

[0005] The purpose of the present invention is to make up for the deficiencies of the existing technology and provide an electric motorcycle lithium-ion battery management system and method. It can realize real-time and accurate monitoring and management of battery status by integrating data acquisition, digital twin model construction, simulation analysis, and strategy formulation modules. It can not only more accurately predict the battery life attenuation but also formulate scientific and reasonable maintenance and replacement strategies according to the prediction results, thereby significantly improving the accuracy and efficiency of battery management. At the same time, the human-machine interaction module equipped with the system provides an intuitive and convenient information display and warning mechanism, enhancing the user experience and the safety of battery use.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: On the one hand, an electric motorcycle lithium-ion battery management system, which includes the following components: a data acquisition module, a digital twin model construction module, a simulation analysis module, and a strategy formulation module;

[0007] The data acquisition module: It is set on the lithium-ion battery of the electric motorcycle and is used to collect the voltage, current, and temperature operation data of the battery in real time, and transmit the collected data to the digital twin model construction module;

[0008] The digital twin model construction module: It receives the data transmitted by the data acquisition module and constructs a digital model in the virtual space that is exactly the same as the state of the physical battery;

[0009] The simulation analysis module: It uses the constructed digital twin model to simulate the performance changes of the battery under different working conditions and environments. Through the analysis of the simulation data, it predicts the battery life attenuation situation, including the change trends of capacity attenuation and internal resistance increase indicators;

[0010] The strategy formulation module: According to the prediction results of the simulation analysis module, combined with the actual use situation of the electric motorcycle and user requirements, it formulates accurate battery maintenance and replacement strategies. When it is predicted that the battery capacity decays to a certain extent, it recommends that the user perform battery maintenance or replacement, and at the same time provides maintenance plans and replacement time suggestions.

[0011] Further, the digital twin model construction module constructs a digital model using an improved adaptive weight particle swarm optimization-neural network hybrid algorithm. The specific algorithm is as follows: First, define the position vector X of the particle swarm i =(x i1 ,x i2 ,…,x in ), the velocity vector V i =(v i1 ,v i2 ,…,v in ), where i = 1, 2, …, m, m is the number of particles, n is the dimension of the model parameters, and the fitness function F(X i ) of the particle is calculated according to the mean square error between the actual battery data and the model prediction data, that is where N is the number of data samples, y j is the actually measured battery performance data, is the battery performance data predicted by the model;

[0012] In each iteration, the velocity update formula of the particle is: v ij (t + 1)=ωv ij (t)+c1r 1j (t)(p ij -x ij (t))+c2r 2j (t)(g j -x ij(t)), where ω is the inertia weight, dynamically adjusted according to the number of iterations by a fuzzy logic system, with an initial value set to 0.9, changing between 0.4 - 0.9 as the number of iterations increases, c1 and c2 are learning factors, both with an initial value set to 2, and are adaptively adjusted based on the difference between the historical optimal solution and the current solution, r 1j and r 2j are random numbers between [0, 1], p ij is the individual historical optimal position of particle i, g j is the global optimal position, and the position update formula for the particle is: x ij (t + 1) = x ij (t) + v ij (t + 1) uses the optimal parameters obtained by the particle swarm optimization algorithm as the initial weights and thresholds of the neural network. The neural network adopts a multi-layer perceptron structure. The number of input layer nodes is determined according to the dimension of the collected battery data, and the number of output layer nodes is the number of battery key performance indicators. The training of the neural network uses an improved stochastic gradient descent algorithm, and the learning rate η is adaptively adjusted according to the change rate of the gradient. The adjustment formula is where η0 is the initial learning rate, set to 0.01, and α is the adjustment coefficient, set to 0.1, is the difference between the current gradient and the previous gradient, is the current gradient. Through the above algorithm, the high-precision construction and dynamic update of the digital twin model are realized. The determination of parameters and weights is to find the optimal solution in the search space through the particle swarm optimization algorithm and for the neural network to train and learn based on actual data, ensuring that the model can accurately reflect the state of the physical battery.

[0013] Furthermore, when the simulation analysis module simulates the performance changes of the battery under different working conditions and environments, it uses a working condition - environment coupling influence prediction algorithm. This algorithm first establishes a working condition feature vector C = (c1, c2,..., c k ), where c i represents different working condition characteristic parameters, and an environment feature vector E = (e1, e2,..., e l ), where e j represents different environment characteristic parameters;

[0014] Then it constructs a coupling influence coefficient matrix M. The matrix element m ij represents the coupling influence degree of the working condition characteristic parameter c i and the environment characteristic parameter e j on the battery performance. The determination method of m ij is: by analyzing a large amount of historical experimental data, using the principal component analysis method to extract the main factors affecting the battery performance, and then combining regression analysis to establish the functional relationship between m ij and the main influencing factors, that is, mij = f(a1, a2, …, a s ), where a k are the main factors affecting battery performance;

[0015] The battery performance prediction formula is where β i is the individual influence weight of the operating condition characteristic parameters, γ j is the individual influence weight of the environmental characteristic parameters, δ is the constant term, and β i , γ j and δ are obtained by fitting historical data through the least squares method. Through this algorithm, the coupling effect of the operating conditions and the environment can be comprehensively considered, and the performance changes of the battery under different conditions can be accurately predicted, providing more reliable data support for the prediction of battery life attenuation.

[0016] Furthermore, when formulating the battery maintenance and replacement strategy, the strategy formulation module adopts a decision-making algorithm based on risk assessment. This algorithm first defines the battery performance risk index R, which is comprehensively calculated from the battery capacity decay rate r c , the internal resistance growth rate r r , and the temperature anomaly frequency f t . The calculation formula is R = ω1r c + ω2r r + ω3f t , where ω1, ω2, and ω3 are the weights of each factor. The method for determining the weights is as follows: By combining the expert scoring method with the analytic hierarchy process, the influence degree of different factors on battery performance and safety is evaluated, a judgment matrix is constructed, and the weights of each factor are calculated;

[0017] Then, according to the magnitude of the risk index R, the risk levels are divided, and corresponding maintenance and replacement strategies are formulated for different risk levels. When R is in the low-risk level, it is recommended to conduct regular battery inspections and maintenance. When R is in the medium-risk level, it is recommended to conduct in-depth battery performance tests and formulate personalized maintenance plans based on the test results. When R is in the high-risk level, it is recommended to immediately replace the battery. At the same time, this algorithm also considers the usage frequency and driving mileage factors of the electric motorcycle to dynamically adjust the maintenance and replacement strategy to ensure the accuracy and practicality of the strategy.

[0018] Furthermore, the data acquisition module uses a multi-sensor data fusion algorithm to process the collected voltage, current, and temperature data. This algorithm first establishes an error model for the sensor data. For the sensor s i , the relationship between its measured value x i and the true value x is x i = x + ∈ i , where ∈ iis the measurement error, which follows a normal distribution is the sensor s i variance;

[0019] Then, the weighted least squares method is used for data fusion, and the fused data The calculation formula is where n is the number of sensors. To further improve the accuracy of data fusion, an adaptive weight adjustment mechanism is introduced, and the weight w i is dynamically adjusted according to the stability of the historical measurement data of the sensor, and the adjustment formula is where is the average variance of the historical measurement data of the sensor s i Through this multi-sensor data fusion algorithm, the accuracy and reliability of the collected data can be effectively improved, providing a high-quality data basis for the subsequent construction and simulation analysis of the digital twin model.

[0020] Furthermore, in the improved adaptive weight particle swarm optimization-neural network hybrid algorithm of the digital twin model construction module, in order to prevent the particle swarm from falling into a local optimum, a chaotic mutation strategy is introduced. The specific process is as follows: During the particle swarm iteration process, when the optimal solution has not changed for T consecutive iterations, a chaotic mutation operation is performed on the current global optimal position. Logistic mapping is used for chaotic initialization, and the formula is z k+1 = μz k (1 - z k ), where μ is the control parameter, and its value range is (0, 4]. Here, μ = 3.9, z k represents the chaotic sequence value generated in the k-th iteration, z0 is a random number between [0, 1], and z k+1 represents the chaotic sequence value generated in the (k + 1)-th iteration. A chaotic sequence Z = (z1, z2,..., z n ) is generated through chaotic mapping, and then a mutation operation is performed on the global optimal position g = (g1, g2,..., g n ). The formula for the mutated position g' = (g'1, g'2,..., g' n ) is g' i = g i + λ(z i - 0.5), where λ is the mutation coefficient, and g i represents the new position value obtained after performing a chaotic mutation operation on the i-th dimension of the global optimal position g = (g1, g2,..., g n ), g i is the original value of the i-th dimension of the global optimal position g, and z i is the chaotic sequence Z = (z1, z2,..., z n) The i-th value in is dynamically adjusted according to the number of iterations. The initial value is set to 0.1 and gradually decreases as the number of iterations increases. By introducing a chaotic mutation strategy, the diversity of the particle swarm in the search space is increased, and the probability of the algorithm finding the global optimal solution is improved, thereby enhancing the construction accuracy of the digital twin model.

[0021] Furthermore, in the working condition-environment coupling impact prediction algorithm of the simulation analysis module, in order to adapt to the change of battery performance over time, a time series correction factor is introduced. Let the battery usage time be t, and the calculation formula of the time series correction factor θ(t) is where α1 and α2 are correction coefficients obtained by fitting the battery aging experimental data, and τ1 and τ2 are time constants determined according to the type and usage characteristics of the battery. The time series correction factor θ(t) is introduced into the battery performance prediction formula to obtain the corrected prediction formula P′ = θ(t) × P. By introducing the time series correction factor, the performance change of the battery at different usage stages can be more accurately reflected, and the accuracy of battery life attenuation prediction can be improved.

[0022] Furthermore, the system further includes a data storage module for storing the collected battery operation data, the constructed digital twin model data, the simulation analysis result data, and the formulated maintenance and replacement strategy data. The data storage module adopts a distributed storage architecture and combines blockchain technology to store the data on multiple nodes. The security and immutability of the data are ensured through the hash algorithm and consensus mechanism of the blockchain. At the same time, the data storage module sets a data regular cleaning and backup strategy. According to the importance and usage frequency of the data, the expired data is cleaned, and the key data is regularly backed up to ensure the efficiency and reliability of the system data storage and provide data guarantee for the long-term stable operation of the system.

[0023] Furthermore, the system further includes a human-machine interaction module for realizing the information interaction between the user and the system. The human-machine interaction module adopts a graphical interface design. The user can view the battery operation status, performance prediction results, and maintenance and replacement strategy suggestions in real time through the interface. At the same time, the user can also input the usage requirements and personalized setting information of the electric motorcycle through the interface, and the system adjusts the battery management strategy according to the information input by the user. The human-machine interaction module also has a voice interaction function, supporting voice query and voice command input to improve the convenience of user operation. In addition, the human-machine interaction module sets a warning prompt function. When the battery has an abnormal situation or reaches the maintenance and replacement threshold, a warning is sent to the user in a timely manner through sound and light to enhance the user's perception ability of the battery status.

[0024] On the other hand, an electric motorcycle lithium-ion battery management method is characterized in that the specific steps of the method are as follows:

[0025] S1. Data acquisition step: The data acquisition module continuously acquires the operating data of the voltage, current, and temperature of the lithium-ion battery of the electric motorcycle and transmits the data to the digital twin model construction module;

[0026] S2. Digital twin model construction step: The digital twin model construction module constructs a digital model consistent with the state of the physical battery in the virtual space according to the received data and continuously updates the model parameters according to the real-time data;

[0027] S3. Simulation analysis step: The simulation analysis module uses the digital twin model to set different working conditions and environmental parameters, simulates the performance changes of the battery under these conditions, analyzes the simulation data, and predicts the battery life attenuation;

[0028] S4. Strategy formulation step: The strategy formulation module formulates precise battery maintenance and replacement strategies according to the prediction results of the simulation analysis module and feeds the strategies back to the user or the relevant management system.

[0029] Compared with the prior art, the lithium-ion battery management system and method for the electric motorcycle have the following beneficial effects:

[0030] First, by integrating modules such as data acquisition, digital twin model construction, simulation analysis, and strategy formulation, the system can continuously acquire the battery operating data, construct a digital model consistent with the state of the physical battery in the virtual space, and use this model to simulate the performance changes of the battery under different working conditions and environments, so as to accurately predict the battery life attenuation. This provides precise battery maintenance and replacement strategy suggestions for users, significantly improving the accuracy and efficiency of battery management.

[0031] Second, the invention is equipped with a human-computer interaction module with a graphical interface design and supports voice interaction functions, enabling users to conveniently view the battery operating status, performance prediction results, and maintenance and replacement strategy suggestions. In addition, when the battery has an abnormal situation or reaches the maintenance and replacement threshold, the system will promptly issue a warning to the user through sounds, lights, etc., thus enhancing the user experience and the safety of battery use.

[0032] Other advantages, objectives, and features of the present invention will be described to some extent in the subsequent specification, and to some extent, will be obvious to those skilled in the art based on the study of the following text, or can be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0034] Figure 1 It is a flow operation diagram of the lithium-ion battery management system for an electric motorcycle;

[0035] Figure 2 It is a flow operation diagram of the lithium-ion battery management method for an electric motorcycle. Specific embodiments

[0036] To further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following will, in conjunction with the drawings and preferred embodiments, detail the specific embodiments, structures, features, and their effects of the present invention as follows.

[0037] Embodiment 1

[0038] This embodiment describes that Xiao Li commutes in the city by electric motorcycle every day, and the driving route includes congested sections and smooth sections, and the environmental temperature changes significantly with the seasons.

[0039] The data acquisition module uses a multi-sensor data fusion algorithm. Assuming that 3 voltage sensors, 3 current sensors, and 3 temperature sensors are used to collect the operating data of the battery, an error model of the sensor data is established. For sensor s i , the measured value x i and the true value x relationship is x i =x + ∈ i , ∈ i obeys a normal distribution The weighted least squares method is used for data fusion, and the fused data and the weights are dynamically adjusted according to the stability of the historical measurement data of the sensors The collected data is transmitted to the digital twin model construction module.

[0040] The digital twin model construction module uses an improved adaptive weight particle swarm optimization-neural network hybrid algorithm, defines the particle swarm position vector X i =(x i1 , x i2 , …, x in ), the velocity vector V i =(v i1 , v i2 , …, v in ), the particle fitness function In each iteration, the particle velocity update formula is v ij (t + 1) = ωv ij (t) + c1r 1j (t)(p ij -x ij (t)) + c2r 2j (t)(g j -x ij (t)), and the position update formula is x ij (t + 1) = x ij (t) + v ij (t + 1). The optimal parameters obtained by the particle swarm optimization algorithm are used as the initial weights and thresholds of the neural network. The neural network training adopts an improved stochastic gradient descent algorithm, and the learning rate The model continuously updates the parameters according to the real-time data.

[0041] The simulation analysis module adopts a working condition - environment coupling impact prediction algorithm, establishes a working condition feature vector C = (c1, c2), where c1 represents the proportion of the driving time in the congested section, and c2 represents the proportion of the driving time in the smooth section; the environmental feature vector E = (e1), representing the environmental temperature, constructs a coupling impact coefficient matrix M, and the element m ij is determined by performing principal component analysis and regression analysis on a large amount of historical experimental data. For example, m 11 = f(a1, a2), and the battery performance prediction formula is Considering the time series correction factor The corrected prediction formula is P' = θ(t)×P. Through simulation analysis, the life attenuation of the battery under different seasons and road conditions is predicted.

[0042] The strategy formulation module adopts a decision-making algorithm based on risk assessment, defines the battery performance risk index R = ω1r c + ω2r r + ω3f t . According to the risk index R, the risk levels are divided. When R is in the low risk level, it is recommended that Xiao Li conduct regular battery inspections and maintenance. When R is in the medium risk level, it is recommended to conduct in-depth battery performance detection and formulate a personalized maintenance plan according to the detection results. When R is in the high risk level, it is recommended to immediately replace the battery. At the same time, considering Xiao Li's usage frequency and driving mileage, the maintenance and replacement strategies are dynamically adjusted.

[0043] Embodiment 2

[0044] This embodiment describes that Xiao Wang is a food delivery rider active in the urban area. His daily working hours exceed 10 hours, and the riding mileage is usually about 100 - 150 kilometers. During work, he frequently shuttles through the streets and alleys, sometimes encountering congestion and slow traffic, and sometimes accelerating on open roads. The operations of sudden acceleration and sudden deceleration are extremely frequent. Moreover, his working hours cover the whole day, experiencing environmental temperature changes at different times. The high temperature in summer can reach above 35°C, and the low temperature in winter may drop to about 0°C. At the same time, he also needs to cope with different weather conditions, such as rainy days and sunny days.

[0045] The data acquisition module relies on high-precision voltage, current, and temperature sensors to monitor the battery operation status in real time. Considering that heavy-season high temperature, winter low temperature, and frequent vibrations may cause fluctuations in sensor measurement errors, a multi-sensor data fusion algorithm is used to ensure data accuracy. Taking the temperature sensor as an example, if the measurement error variance increases for a certain temperature sensor, according to the adaptive weight adjustment mechanism its weight in data fusion will be correspondingly reduced, thereby weakening the adverse impact of this sensor on the overall data accuracy. The data after being fused by the weighted least squares method can more accurately reflect the actual state of the battery and be transmitted to the digital twin model construction module in real time.

[0046] The digital twin model construction module uses an improved adaptive weight particle swarm optimization-neural network hybrid algorithm to accurately construct a digital model that matches the state of the physical battery. During the particle swarm iteration process, closely monitor the change of the optimal solution. Once the optimal solution has not changed for consecutive T iterations, immediately start the chaotic mutation strategy.

[0047] Use the Logistic mapping z k+1 = μz k (1 - z k ) for chaotic initialization, where μ is taken as 3.8 (after multiple tests, this value can effectively enhance the algorithm's optimization ability in the current scenario), z0 is a random number between [0, 1], and a chaotic sequence Z = (z1, z2,..., z n ) is generated through this mapping. Then, perform mutation operations on the global optimal position g = (g1, g2,..., g n ). The mutated position g' i = g i + λ(z i - 0.5), where λ is set to 0.2 (based on actual tests, this mutation coefficient can avoid the algorithm from falling into local optimality while ensuring the model convergence speed). The optimized particle swarm parameters are used as the initial weights and thresholds of the neural network, and combined with the improved stochastic gradient descent algorithm to train the neural network. The learning rate Let η0 be 0.01 and α be set to 0.5 to ensure that the model can reflect the changes in battery status in real time and accurately.

[0048] The simulation analysis module adopts a working condition - environment coupling impact prediction algorithm, comprehensively considering the impacts of the complex working conditions and variable environments of food delivery on battery performance, constructing a working condition feature vector C = (c1, c2, c3), where c1 represents the daily driving mileage, c2 represents the number of rapid accelerations and decelerations, c3 is the single - ride duration, and an environment feature vector E = (e1, e2), where e1 represents the environmental temperature and e2 represents the air humidity.

[0049] By performing principal component analysis on a large amount of historical experimental data, the main factors affecting battery performance are screened out, such as the battery chemical reaction rate and internal resistance change, and the elements m in the coupling impact coefficient matrix M are determined by combining regression analysis. ij For example, m 11 = 0.3a1 + 0.5a2 - 0.2 (a1 and a2 are the main factors affecting battery performance).

[0050] The battery performance prediction formula is β i 、γ j and δ are obtained by fitting historical data. At the same time, a time - series correction factor α1 = 0.1, α2 = 0.05, τ1 = 30 (days), τ2 = 365 (days), and the corrected prediction formula is P′ = θ(t)×P.

[0051] Simulate different delivery scenarios, such as high - temperature periods in summer, low - temperature periods in winter, and high - humidity environments in rainy days, to predict life - decay indicators such as battery capacity decay and internal resistance increase, providing a reliable basis for subsequent strategy formulation.

[0052] The strategy - making module, based on a decision - making algorithm for risk assessment and combined with the characteristics of the food - delivery business, formulates scientific and reasonable battery maintenance and replacement strategies, defining a battery performance risk index R = ω1r c + ω2r r + ω3f t According to practical experience, set ω1 = 0.4, ω2 = 0.3, ω3 = 0.3.

[0053] Taking into account the high frequency and long mileage characteristics of Xiao Wang's work, targeted adjustments are made to the risk level classification standards. When R is at a low risk level, it is recommended to conduct routine battery inspection and maintenance once a week, such as cleaning the battery surface and checking the connection lines. When R is at a medium risk level, in view of the timeliness of delivery work, during the next shift change, Xiao Wang is arranged to send the vehicle to a designated maintenance site for in-depth battery performance testing, and customize a personalized maintenance plan based on the test results, such as balanced charging, replacement of some aging components, etc. When R is at a high risk level, the vehicle is immediately decommissioned and replaced with a spare battery to ensure that the delivery work is not affected, and the faulty battery is repaired or scrapped in a timely manner.

[0054] In addition, based on Xiao Wang’s real-time delivery data, such as the mileage traveled on the day and the estimated number of remaining orders, the maintenance and replacement strategies are dynamically adjusted to extend the battery life and reduce operating costs while ensuring delivery efficiency.

[0055] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. The lithium-ion battery management system for an electric motorcycle is characterized in that, The system includes the following components: a data acquisition module, a digital twin model construction module, a simulation analysis module, and a strategy formulation module: The data acquisition module: It is set on the lithium-ion battery of the electric motorcycle and is used to collect the voltage, current, and temperature operation data of the battery in real time, and transmit the collected data to the digital twin model construction module; The digital twin model construction module: Receives the data transmitted by the data acquisition module and constructs a digital model in the virtual space that is exactly the same as the state of the physical battery; The simulation analysis module: Utilizes the constructed digital twin model to simulate the performance changes of the battery under different working conditions and environments, and predicts the battery life attenuation situation through the analysis of the simulation data, including the change trends of capacity attenuation and internal resistance increase indicators; The strategy formulation module: According to the prediction results of the simulation analysis module, combined with the actual usage situation of the electric motorcycle and user requirements, formulates precise battery maintenance and replacement strategies. When it is predicted that the battery capacity decays to a certain extent, it recommends that the user perform battery maintenance or replacement, and at the same time provides a maintenance plan and replacement time suggestion.

2. The lithium-ion battery management system for an electric motorcycle according to claim 1, wherein The digital twin model construction module constructs a digital model using an improved adaptive weight particle swarm optimization-neural network hybrid algorithm. The specific algorithm is as follows: First, define the position vector X of the particle swarm i =(x i1 , x i2 , …, x in ), the velocity vector V i =(v i1 , v i2 , …, v in ), where i = 1, 2, …, m, m is the number of particles, n is the dimension of the model parameters, and the fitness function F(X i ) of the particle is calculated according to the mean square error between the actual battery data and the model prediction data, that is where N is the number of data samples, y j is the actually measured battery performance data, is the battery performance data predicted by the model; In each iteration, the velocity update formula of the particle is: v ij (t + 1) = ωv ij (t) + c1r 1j (t)(p ij - x ij (t)) + c2r 2j (t)(g j - x ij (t)), where ω is the inertia weight, which is dynamically adjusted according to the number of iterations through a fuzzy logic system, c1 and c2 are learning factors, r 1j and r 2j are random numbers between [0, 1], p ij is the individual historical optimal position of particle i, g j is the global optimal position. The position update formula of the particle is: x ij (t + 1)) = x ij (t) + v ij (t + 1). The optimal parameters obtained by the particle swarm optimization algorithm are used as the initial weights and thresholds of the neural network. The neural network adopts a multi-layer perceptron structure, and the number of output layer nodes is the number of battery key performance indicators. The training of the neural network uses an improved stochastic gradient descent algorithm, and the learning rate η is adaptively adjusted according to the change rate of the gradient. The adjustment formula is where η0 is the initial learning rate, α is the adjustment coefficient, is the difference between the current gradient and the previous gradient, is the current gradient.

3. The lithium-ion battery management system for an electric motorcycle according to claim 1, characterized in that, When the simulation analysis module simulates the performance changes of the battery under different working conditions and environments, it uses a working condition-environment coupling impact prediction algorithm. This algorithm first establishes a working condition feature vector C = (c1, c2, …, c k ), where c i represents different working condition characteristic parameters, and an environment feature vector E = (e1, e2, …, e l ), where e j represents different environment characteristic parameters; Then, a coupling influence coefficient matrix M is constructed, and the matrix element m ij represents the coupling influence degree of the operating condition characteristic parameter c i and the environmental characteristic parameter e j on the battery performance. The determination method of m ij is as follows: By analyzing a large amount of historical experimental data, the main factors affecting the battery performance are extracted using the principal component analysis method, and then a function relationship between m ij and the main influencing factors is established through regression analysis, that is, m ij = f(a1, a2, …, a s ), where a k are the main factors affecting the battery performance; The battery performance prediction formula is where β i is the individual influence weight of the operating condition characteristic parameters, γ j is the individual influence weight of the environmental characteristic parameters, and δ is the constant term.

4. The lithium-ion battery management system for an electric motorcycle according to claim 1, characterized in that, When formulating the battery maintenance and replacement strategy, the strategy formulation module adopts a decision-making algorithm based on risk assessment. This algorithm first defines the battery performance risk index R, which is comprehensively calculated from the battery capacity attenuation rate r c , the internal resistance growth rate r r , and the abnormal temperature frequency f t . The calculation formula is R = ω1r c + ω2r r + ω3f t , where ω1, ω2, and ω3 are the weights of each factor; Then, the risk level is divided according to the magnitude of the risk index R, and corresponding maintenance and replacement strategies are formulated for different risk levels. When R is at a low risk level, it is recommended to conduct regular battery inspections and maintenance. When R is at a medium risk level, it is recommended to conduct in-depth battery performance tests, and a personalized maintenance plan is formulated according to the test results. When R is at a high risk level, it is recommended to immediately replace the battery. At the same time, this algorithm also considers the usage frequency and mileage factors of the electric motorcycle to dynamically adjust the maintenance and replacement strategies.

5. The lithium-ion battery management system for an electric motorcycle according to claim 1, characterized in that The data acquisition module processes the collected voltage, current, and temperature data using a multi-sensor data fusion algorithm. This algorithm first establishes an error model for the sensor data. For sensor s i , its measured value x i and the relationship with the true value x is x i = x + ∈ i , where ∈ i is the measurement error and follows a normal distribution is the variance of sensor s i . Then, weighted least squares method is used for data fusion, and the fused data The calculation formula is where n is the number of sensors, and an adaptive weight adjustment mechanism is introduced. The weight w i is dynamically adjusted according to the stability of the historical measurement data of the sensors. The adjustment formula is where is the average variance of the historical measurement data of sensor s i ​ 6. The lithium-ion battery management system for an electric motorcycle according to claim 2, characterized in that, In the improved adaptive weight particle swarm optimization-neural network hybrid algorithm of the digital twin model construction module, a chaotic mutation strategy is introduced. The specific process is as follows: During the particle swarm iteration process, when the optimal solution has not changed for consecutive T iterations, a chaotic mutation operation is performed on the current global optimal position. Logistic mapping is used for chaotic initialization, and the formula is z k+1 = μz k (1 - z k ), where μ is the control parameter, z k represents the chaotic sequence value generated in the k-th iteration, z0 is a random number between [0, 1], z k+1 represents the chaotic sequence value generated in the (k + 1)-th iteration. A chaotic sequence Z = (z1, z2,..., z n ) is generated through chaotic mapping. Then, a mutation operation is performed on the global optimal position g = (g1, g2,..., g n ). The formula for the mutated position g' = (g'1, g'2,..., g' n ) is g' i = g i + λ(z i - 0.5), where λ is the mutation coefficient, g' i represents the new position value obtained after performing a chaotic mutation operation on the i-th dimension of the global optimal position g = (g1, g2,..., g n ), g i is the original value of the i-th dimension of the global optimal position g, and z i is the i-th value in the chaotic sequence Z = (z1, z2,..., z n ) generated through Logistic mapping.

7. The lithium-ion battery management system for an electric motorcycle according to claim 3, wherein, In the working condition-environment coupling influence prediction algorithm of the simulation analysis module, a time series correction factor is introduced. Let the battery usage time be t, and the calculation formula of the time series correction factor θ(t) is where α1 and α2 are correction coefficients, τ1 and τ2 are time constants. The time series correction factor θ(t) is introduced into the battery performance prediction formula to obtain the corrected prediction formula P′ = θ(t) × P.

8. The lithium-ion battery management system for an electric motorcycle according to claim 1, wherein The system further includes a data storage module, which is used to store the collected battery operation data, the constructed digital twin model data, the simulation analysis result data, and the formulated maintenance and replacement strategy data. The data storage module adopts a distributed storage architecture and combines blockchain technology to store the data on multiple nodes, and ensures the security and immutability of the data through the blockchain's hash algorithm and consensus mechanism. At the same time, the data storage module sets a data regular cleaning and backup strategy, and cleans the expired data and backs up the key data regularly according to the importance and usage frequency of the data.

9. The lithium-ion battery management system for an electric motorcycle according to claim 1, wherein The system further includes a human-computer interaction module, which is used to realize the information interaction between the user and the system. The human-computer interaction module adopts a graphical interface design. The user can view the battery operation status, performance prediction results, and maintenance and replacement strategy suggestions in real time through the interface. At the same time, the user can also input the usage requirements and personalized setting information of the electric motorcycle through the interface, and the system adjusts the battery management strategy according to the information input by the user. The human-computer interaction module also has a voice interaction function, supporting voice queries and voice command inputs. In addition, the human-computer interaction module sets a warning prompt function, and when the battery has an abnormal situation or reaches the maintenance and replacement threshold, it issues a warning to the user in a timely manner through sound and light.

10. The lithium-ion battery management method for an electric motorcycle is applicable to the lithium-ion battery management system of the electric motorcycle described in any one of claims 1-9, and is characterized in that, The specific steps of this method are as follows: S1. Data acquisition step: The data acquisition module continuously collects the voltage, current, and temperature operation data of the lithium-ion battery of the electric motorcycle and transmits the data to the digital twin model construction module; S2. Digital twin model construction step: The digital twin model construction module constructs a digital model consistent with the state of the physical battery in the virtual space based on the received data and continuously updates the model parameters according to the real-time data; S3. Simulation analysis step: The simulation analysis module uses the digital twin model to set different working conditions and environmental parameters, simulates the performance changes of the battery under these conditions, analyzes the simulation data, and predicts the battery life attenuation; S4. Strategy formulation step: The strategy formulation module formulates precise battery maintenance and replacement strategies based on the prediction results of the simulation analysis module and feeds the strategies back to the user or the relevant management system.

Images