Battery charging control method and system based on artificial intelligence

Abstract

The embodiment of the invention provides a battery charging control method and system based on artificial intelligence, and the method comprises the steps: obtaining the battery state data in the operation process of a battery, and enabling the battery state data to comprise the voltage state data, current state data, temperature state data and impedance state data of the battery; and calculating the SOH (state of health) of the battery. And according to the battery state data and the SOH, constructing a target feature sequence, and inputting the target feature sequence into a pre-trained residual service life prediction model to obtain a residual service life prediction result of the battery. And adjusting a charging strategy of the battery according to the SOH of the battery and the residual service life prediction result to obtain a target charging strategy, and charging the battery based on the target charging strategy, thereby improving the stability of the battery system and prolonging the service life of the battery system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of battery, in particular to a battery charging control method and system based on artificial intelligence. BACKGROUND

[0002] Key infrastructures such as communication base stations and data centers have put forward very high stability and life requirements for power supply systems. Lithium ion batteries are widely used in backup power supply systems due to their high energy density and excellent cycle performance. However, current charging management is still generally carried out in a constant voltage and current limiting (CV / CC) mode, ignoring the aging characteristics of the battery in a variable environment and during long-term operation. The traditional battery management system (BMS) mainly uses a simple temperature and voltage threshold judgment method, which cannot perform fine prediction and dynamic control. This not only increases the potential failure risk of the battery system, but also limits the further improvement of the cycle life of the battery system.

[0003] Therefore, how to improve the stability and life of the battery system is a problem to be solved. SUMMARY

[0004] Therefore, the purpose of the present application is to provide a battery charging control method and system based on artificial intelligence to improve the stability and life of the battery system.

[0005] In a first aspect, the present application provides a battery charging control method based on artificial intelligence, comprising: obtaining battery state data in the operation of the battery, the battery state data including voltage state data, current state data, temperature state data and impedance state data of the battery; calculating the state of health (SOH) of the battery; constructing a target feature sequence according to the battery state data and the SOH, and inputting the target feature sequence into a pre-trained remaining useful life prediction model to obtain a remaining useful life prediction result of the battery; adjusting the charging strategy of the battery according to the SOH of the battery and the remaining useful life prediction result to obtain a target charging strategy; charging the battery based on the target charging strategy.

[0006] Optionally, the calculation of the SOH of the battery comprises: obtaining a nominal capacity of the battery; calculating the SOH of the battery according to the nominal capacity of the battery, the temperature state data, the current state data and a preset parameter, the preset parameter including a preset weight coefficient, an activation energy constant and a gas constant.

[0007] Optionally, constructing a target feature sequence according to the battery state data and the SOH, and inputting the target feature sequence into a pre-trained remaining useful life prediction model to obtain a remaining useful life prediction result of the battery includes: Extracting local features of the feature sequence through a one-dimensional convolutional neural network in the remaining useful life prediction model; Extracting temporal dependency features in the feature sequence through a bidirectional gated recurrent unit network in the remaining useful life prediction model; Determining feature weights through an attention mechanism in the remaining useful life prediction model; The remaining service life prediction result of the battery is generated according to the local features, the timing dependency features, and the feature weights.

[0008] Optionally, adjusting the charging strategy of the battery according to the SOH of the battery and the remaining service life prediction result to obtain a target charging strategy includes: If the remaining service life prediction result is less than or equal to the preset remaining service life, or the SOH is less than or equal to the preset SOH, obtaining a target charging current according to the SOH, the SOH health threshold, and the SOH change rate; Based on the target charging current, obtaining a target charging output duty cycle; The target charging strategy is determined according to the target charging current and the target charging output duty cycle.

[0009] Optionally, obtaining a target charging current according to the SOH, the SOH health threshold, and the SOH change rate includes: Obtaining a first experience parameter and a second experience parameter; Calculating a quotient of the SOH and the SOH health threshold; The target charging current is calculated according to the first empirical parameter, the second empirical parameter, a quotient of the SOH and the SOH health threshold, and the SOH change rate.

[0010] Optionally, if the remaining useful life prediction result is less than or equal to a preset remaining useful life, or the SOH is less than or equal to a preset SOH, the method further includes: Output charging strategy prompt information, where the charging strategy prompt information is used to prompt the user that the battery will switch to a life-extending slow charging mode.

[0011] Optionally, after charging the battery based on the target charging strategy, the method further includes: obtaining the core temperature and heat dissipation point temperature of the battery during the charging process; If the difference between the core temperature and the heat dissipation point temperature is greater than or equal to a preset temperature fluctuation threshold, the amount of heat to be controlled is calculated based on the core temperature and the heat dissipation point temperature; Temperature control is performed on the charging process of the battery according to the heat to be controlled.

[0012] Optionally, the method further includes: Obtaining the actual remaining service life of the battery; Obtaining a model optimization objective function of the remaining useful life prediction model according to the actual remaining useful life, the remaining useful life prediction result, the self-learning weight parameter set, and the regularization term weight factor; The remaining useful life prediction model is optimized based on the model optimization objective function.

[0013] Optionally, obtaining a model optimization objective function of the remaining useful life prediction model according to the actual remaining useful life, the remaining useful life prediction result, the self-learning weight parameter set, and the regularization term weight factor includes: Calculating the error between the actual remaining useful life and the remaining useful life prediction result; Adjusting the self-learning weight parameter set by the regularization term weight factor to obtain an adjusted self-learning weight parameter set; Based on the error result and the adjusted self-learning weight parameter set, the model optimization objective function of the remaining useful life prediction model is obtained.

[0014] In a second aspect, the present application provides an artificial intelligence-based battery charging control system, which includes a machine-readable storage medium and a processor, wherein the machine-readable storage medium stores machine-executable instructions. When the processor executes the machine-executable instructions, the artificial intelligence-based battery charging control system implements the aforementioned artificial intelligence-based battery charging control method.

[0015] The artificial intelligence-based battery charging control method and system provided herein obtain battery status data during battery operation, including voltage, current, temperature, and impedance data. The system calculates the battery's state of health (SOH). A target feature sequence is constructed based on the battery status data and SOH, and the target feature sequence is input into a pre-trained remaining useful life prediction model to obtain a predicted remaining useful life (RUS) result. Based on the SOH and the RUS prediction result, the battery's charging strategy is adjusted to obtain a target charging strategy. The battery is then charged based on the target charging strategy. By dynamically adapting the upper and lower charging voltage limits, current intensity, and charging duration, the charging process consistently matches the battery's current health and expected lifespan. This avoids the risk of overcharging caused by fixed parameters in traditional constant voltage and current limiting strategies, or the side reactions and thermal runaway risks exacerbated by high-intensity charging during the battery aging phase. This ensures smoother chemical reactions and a more stable structure throughout the battery's lifespan, reduces unnecessary energy loss and degradation damage, and thus improves the stability and lifespan of the battery system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0017] Figure 1 A flowchart of a battery charging control method based on artificial intelligence provided in an embodiment of the present application; Figure 2 A schematic structural diagram of an artificial intelligence-based battery charging control system provided in an embodiment of the present application.

[0018] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0019] To help those skilled in the art better understand the present invention, the following will provide a clear and complete description of the technical solutions in the embodiments of the present invention, in conjunction with the accompanying drawings. It is apparent that the described embodiments are only a portion of the embodiments of the present invention, not all of them. All other embodiments derived by those skilled in the art based on the embodiments of the present invention without inventive effort are intended to fall within the scope of protection of the present invention.

[0020] The terms "first", "second", and the like in the description and in the claims of the present application and above-described drawings are used to distinguish different objects, and are not used to describe a particular sequential order. Moreover, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, device, product, or the like that comprises a list of steps or units is not limited to the listed steps or units, but can optionally further include steps or units not listed, or can optionally further include other steps or units inherent to such processes, methods, products, or the like.

[0021] Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase that the phrase in various places in the specification are not necessarily all referring to the same embodiment, or are necessarily referring to different or alternative embodiments. It is explicitly understood that the embodiments described herein can be combined with each other, implicitly and explicitly.

[0022] Figure 1 A flowchart of a battery charging control method based on artificial intelligence provided by the embodiment of the present application is shown. It should be understood that in other embodiments, the order of some steps of the battery charging control method based on artificial intelligence of the present embodiment can be shared according to actual needs, or some steps can be omitted or maintained. As shown in the figure, the method can include the following steps: Figure 1 S110, acquiring battery state data in the battery operation process, the battery state data including voltage state data, current state data, temperature state data, and impedance state data of the battery.

[0023] In this step, the battery may, for example, be a battery cluster composed of lithium iron phosphate battery cells or ternary lithium battery cells. The acquisition of battery state data is completed by a data acquisition unit arranged at each string of battery cell loops, output loops, and single body interfaces of the battery module, including temperature sensors, voltage detection chips, current detection resistors, Hall elements, and impedance spectrum modules, etc. Among them, the temperature sensor is used to collect temperature state data T(t), the voltage detection chip collects voltage state data V(t) (including single battery cell voltage and total voltage), the current detection resistor cooperates with the Hall element to collect current state data I(t) (including charging and discharging current), and the impedance spectrum module collects impedance state data Z(t) (including impedance values of multiple frequency points) in the range of 10Hz-10kHz through alternating current excitation. These data will be periodically transmitted to the embedded AI controller.

[0024] ​Exemplarily, the temperature sensors are arranged in a distributed manner, with 8 temperature sensors set at different positions of the battery module to collect temperature status data of different battery cells or module positions respectively. The sampling frequency of each sensor is 10 Hz, so the temperature status data T(t) is an 8-dimensional time series. For example, at time t, T(t) can be expressed as [T1(t), T2(t), T3(t), T4(t), T5(t), T6(t), T7(t), T8(t)], where each element corresponds to the temperature value at a different position, in degrees Celsius.

[0025] The voltage detection chip is used to collect the voltage of each string of battery cells and the total output voltage of the battery module. The sampling frequency is 20Hz. The voltage status data V(t) is a multi-dimensional sequence, including the voltage of each single cell and the total voltage. For example, for a battery module consisting of 16 strings of battery cells, V(t) can be expressed as [V1(t), V2(t), ..., V16(t), Vtotal(t)], in volts, where V1(t) to V16(t) are the voltages of each single cell at time t, and Vtotal(t) is the total voltage of the module.

[0026] The current detection resistor and Hall element work together to collect current status data I(t) at a frequency of 20Hz. The current status data includes charging current and discharging current. When the battery is in the charging state, I(t) is positive and when discharging, it is negative. The unit is ampere. Its data format is a one-dimensional time series. However, due to the continuous sampling time, a multi-time sequence set is formed, such as [I(t-2Δt), I(t-Δt), I(t)], where Δt is the sampling interval, and here Δt=0.05s (corresponding to a 20Hz sampling frequency).

[0027] The impedance spectrum module collects impedance state data Z(t) through AC excitation and can obtain the Z(f) curve in the range of 10 Hz to 10 kHz. The sampling frequency is 1 Hz, and an impedance sequence containing multiple frequency points is collected every second. Therefore, Z(t) is a multidimensional sequence. For example, at time t, Z(t) can be expressed as [Z(10 Hz, t), Z(100 Hz, t), Z(1 kHz, t), Z(5 kHz, t), Z(10 kHz, t)], with units of ohms, including impedance values ​​at different frequencies.

[0028] S120: Calculate the state of health (SOH) of the battery.

[0029] S121. Obtain the nominal capacity of the battery.

[0030] In this step, the battery nominal capacity Cr is the capacity parameter marked on the battery when it leaves the factory. For example, the nominal capacity of a lithium-ion battery commonly used in communication backup power supplies can be set to 100Ah. This parameter is pre-stored in the embedded AI controller as the basic parameter for calculating SOH.

[0031] S122. Calculate the SOH of the battery according to the battery nominal capacity, the temperature state data, the current state data, and preset parameters, where the preset parameters include a preset weight coefficient, an activation energy constant, and a gas constant.

[0032] In this step, the SOH can be calculated as shown in the following formula (1): (1); The integral interval is from the initial time 0 when the battery starts to be used to the current time t, and the preset parameters include the preset weight coefficient α (dimensionless, such as 0.001), the activation energy constant Ea (such as 30000 J / mol), the gas constant |I(τ)| is the absolute value of the current in the current state data, T(τ) is the temperature value in the temperature state data, and the value of SOH is between 0 and 1.

[0033] S130: construct a target feature sequence according to the battery status data and the SOH, and input the target feature sequence into a pre-trained remaining service life prediction model to obtain a remaining service life prediction result of the battery.

[0034] Among them, the remaining service life prediction model can be calculated through the NPU acceleration chip, which supports FP16 tensor operations. The number of model parameters is less than or equal to 3M, and the reasoning of the remaining service life prediction can be completed within 100ms.

[0035] In this embodiment, a target feature sequence is first constructed based on the battery status data and SOH. The battery status data includes voltage status data V(t), current status data I(t), temperature status data T(t), and impedance status data Z(t). Assume that V(t) is a 17-dimensional sequence containing 16 single cell voltages and 1 total voltage. The data from the three most recent moments are taken: V(t-2Δt), V(t-Δt), and V(t). The V data at each moment is 17-dimensional, so the voltage-related features are 3×17=51-dimensional. I(t) is a current sequence. The data from the three most recent moments are taken: I(t-2Δt), I(t-Δt), and I(t), resulting in a 3-dimensional sequence. T(t) is the temperature data from eight temperature sensors. The data at the current moment t is taken, resulting in an 8-dimensional sequence. Z(t) is the impedance data at five frequency points. The data at the current moment t is taken, resulting in a 5-dimensional sequence. The SOH is the currently calculated value of 0.5, resulting in a 1-dimensional sequence. These data are concatenated in sequence to construct a target feature sequence x, whose dimension is 51+3+8+5+1=68 dimensions, that is, x=[V(t-2Δt), V(t-Δt), V(t), I(t-2Δt), I(t-Δt), I(t), T(t), Z(t), SOH], where each element is a corresponding data sequence or value.

[0036] Next, the target feature sequence is input into the pre-trained RSU prediction model. This RSU prediction model uses a 1D-CNN, BiGRU, and Attention combination structure. Its specific structure and working process are as follows: S131. Extracting local features of the feature sequence through a one-dimensional convolutional neural network in the remaining useful life prediction model.

[0037] In this step, the one-dimensional convolutional neural network (1D-CNN) consists of multiple convolutional and pooling layers. The convolutional layers perform convolution operations on the target feature sequence using convolution kernels of different sizes. For example, 32 convolution kernels of size 3 are used to extract the first-level local features, and 64 convolution kernels of size 2 are used to further extract them. The pooling layer reduces the feature dimensionality through maximum pooling, ultimately obtaining local features containing local correlation information of voltage, current, temperature, and impedance.

[0038] For example, a 1D-CNN layer consists of two convolutional layers and one pooling layer. The first convolutional layer has 32 convolution kernels, a kernel size of 3 (processing three adjacent feature dimensions at a time), a stride of 1, and uses the ReLU activation function to convolve the input 68-dimensional target feature sequence, outputting 32 feature maps, each with a dimension of 68-3+1=66. The second convolutional layer has 64 convolution kernels, a kernel size of 2, and a stride of 1. Similarly, the ReLU activation function is used to convolve the 32 66-dimensional feature maps output by the first convolutional layer, outputting 64 feature maps, each with a dimension of 66-2+1=65. Then, a maximum pooling layer with a pooling window size of 2 and a step size of 2 is used to pool the 64 65-dimensional feature maps, outputting 64 32-dimensional (65 / / 2=32) local feature maps. These local feature maps extract local feature information from the target feature sequence, such as the local fluctuation characteristics of voltage and the local correlation characteristics of current and temperature.

[0039] S132. Extracting temporal dependency features in the feature sequence through a bidirectional gated recurrent unit network in the remaining useful life prediction model.

[0040] In this step, the Bidirectional Gated Recurrent Unit (BiGRU) network consists of a two-layer bidirectional structure, with 128 hidden units in each layer. The local feature maps output by the 1D-CNN are flattened and then fed into the BiGRU. The forward layer captures temporal correlations from the past to the present, while the reverse layer captures temporal correlations from the present to the future. After processing these two layers, the output is a concatenated temporal dependency feature vector, which contains long-term correlation information about state data at different times (such as the relationship between current change trends and aging rate).

[0041] For example, the 64 32-dimensional local feature maps output by the 1D-CNN layer are flattened to obtain a 64×32=2048-dimensional feature vector as the input to the BiGRU layer. The first BiGRU layer performs forward and reverse processing on this feature vector. The forward processing captures the temporal dependencies from the past to the present, and the reverse processing captures the temporal dependencies from the present to the future. It outputs two 128-dimensional feature vectors, which are concatenated to obtain a 256-dimensional feature vector. The second BiGRU layer uses the 256-dimensional feature vector output by the first layer as input and also performs forward and reverse processing. It outputs two 128-dimensional feature vectors, which are concatenated to obtain a 256-dimensional temporal dependency feature vector. This vector contains long-term temporal correlation information between data at different times in the target feature sequence, such as the temporal relationship between the current change trend and the battery aging rate, and the impact of temperature fluctuations on impedance over time.

[0042] S133. Determine feature weights through the attention mechanism in the remaining useful life prediction model.

[0043] In this step, the attention mechanism maps the temporal dependency feature vector output by the BiGRU into a weight vector through a fully connected layer. This is then normalized using softmax to obtain the weights of each feature dimension (the sum of the weights is 1). For example, feature dimensions related to SOH changes are given higher weights, while dimensions with less influence are given lower weights, thereby highlighting the impact of key features on remaining useful life prediction.

[0044] Specifically, for example, a 256-dimensional feature vector can be mapped to a 256-dimensional weight vector through a fully connected layer. This weight vector is then normalized using the softmax function to obtain a weight value for each dimension, where the sum of the weight values ​​is 1. For example, dimensions related to SOH changes may be assigned a higher weight (such as 0.1), while dimensions with less influence may be assigned a lower weight (such as 0.001). Each dimension value in the time-dependent feature vector is then multiplied by the corresponding weight value to obtain the weighted time-dependent feature vector.

[0045] S134. Generate the remaining service life prediction result of the battery according to the local features, the timing dependency features, and the feature weights.

[0046] In this step, the weighted time-dependent feature vector is input into the fully connected layer, which outputs the remaining service life prediction values ​​corresponding to multiple future time nodes to form a multi-dimensional remaining service life prediction result (such as the remaining number of cycles, remaining service life duration, etc.).

[0047] S140 : Adjust the charging strategy of the battery according to the state of health (SOH) of the battery and the remaining service life prediction result to obtain a target charging strategy.

[0048] S141. If the remaining service life prediction result is less than or equal to the preset remaining service life, or the SOH is less than or equal to the preset SOH, a target charging current is obtained according to the SOH, the SOH health threshold, and the SOH change rate.

[0049] S1411. Obtain a first empirical parameter and a second empirical parameter.

[0050] In this step, the first empirical parameter k1 and the second empirical parameter k2 are preset adjustable parameters, such as k1=0.5, k2=0.3, etc., which are used to weight the effects of the SOH deviation and the health change rate on the charging current.

[0051] S1412: Calculate the quotient of the SOH and the SOH health threshold.

[0052] In this step, the SOH health threshold SOH_thr is a set value (such as 0.8 or 0.9), and the ratio of the current SOH to SOH_thr is calculated, for example, when SOH = 0.5 and SOH_thr = 0.8, the quotient is 0.5 / 0.8 = 0.625.

[0053] S1413, according to the first experience parameter, the second experience parameter, the quotient of the SOH and the SOH health threshold, the SOH change rate, the target charging current is calculated.

[0054] Wherein, the SOH change rate can be obtained by linear fitting of the SOH value in a preset time period. For example, the SOH change rate can be calculated by linear fitting of the SOH values in the last 10 days, assuming that the SOH values in the last 10 days are 0.55, 0.54, 0.53, 0.52, 0.51, 0.50, 0.49, 0.48, 0.47, 0.46, the SOH change rate is-0.01 / day (decrease by 0.01 per day) by linear fitting.

[0055] In this step, the target charging current can be calculated by the following formula (2): (2) ; Wherein, dSOH / dt is the SOH change rate. For example, k1 = 0.5, 1-SOH / SOH_thr = 0.375, k2 = 0.3, dSOH / dt =-0.01, the target charging current is 0.5 x 0.375 + 0.3 x (-0.01) = 0.1845A.

[0056] S142, based on the target charging current, the target charging output duty cycle is obtained.

[0057] In this step, the target charging output duty cycle D chg can be calculated by the following formula (3): (3) ; Wherein I max is the maximum charging current of the battery (such as 10A). Assuming I chg * = 0.1845A, the target charging output duty cycle D chg = 0.1845 / 10 = 0.01845.

[0058] S143, according to the target charging current, the target charging output duty cycle, the target charging strategy is determined.

[0059] In this step, the target charging strategy includes a target charging current (such as 0.1845A) and a target charging output duty cycle (such as 0.01845). The dynamic charge and discharge execution unit adjusts the PWM signal and current output according to the strategy to achieve charging control that takes into account both charging efficiency and delayed aging.

[0060] S150: Charge the battery based on the target charging strategy.

[0061] In this step, for example, a target charging strategy instruction can be received by a dynamic charge and discharge execution unit (such as a bidirectional DC-DC converter, a MOSFET control unit, etc.), and the switch tube is controlled to be turned on and off by adjusting the duty cycle of the PWM signal, so that the actual charging current is stabilized near the target charging current, thereby completing the charging operation of the battery.

[0062] In one possible implementation, if the remaining service life prediction result is less than or equal to the preset remaining service life, or the SOH is less than or equal to the preset SOH, a charging strategy prompt message may be output, and the charging strategy prompt message is used to prompt the user that the battery will switch to the life-extending slow charging mode. In this implementation, assuming that the preset remaining service life L crit (such as 300 cycles) and preset SOH (SOH crit , such as 0.6, is the judgment threshold. When the trigger condition is met, the embedded AI controller generates a prompt message (such as an electrical signal), transmits it to the monitoring terminal, and displays a text prompt to inform the user that the battery will switch to the life-extending slow charging mode.

[0063] Optionally, after charging the battery based on the target charging strategy, the method may further include the following steps: S210 : Acquire the core temperature and heat dissipation point temperature of the battery during the charging process.

[0064] In this step, the core temperature T_core is the temperature of the battery module's central cell, and the heat sink temperature T_sink is the temperature at the point where the heat sink contacts the outside world. These temperatures are collected in real time by temperature sensors located at corresponding locations, and are measured in degrees Celsius. The heat sink can be installed on the outer surface of the battery module or in a separate layer within the module, and may include, for example, a heat pipe, thermal grease, or a graphite heat spreader.

[0065] S220: If the difference between the core temperature and the heat dissipation point temperature is greater than or equal to a preset temperature fluctuation threshold, calculate the amount of heat to be controlled according to the core temperature and the heat dissipation point temperature.

[0066] In this step, the preset temperature fluctuation threshold is 2°C. When |T_core-T_sink|≥2°C, the heat to be controlled q (in W) is calculated using the heat conduction formula q=λ×A×(T_core-T_sink) / R_th, where λ is the thermal conductivity of the heat dissipation material of the heat dissipation structure (e.g. ), A is the heat transfer area of ​​the heat dissipation structure (such as 0.1m²), and R_th is the thermal resistance (such as 0.1K / W).

[0067] S230: Perform temperature control on the charging process of the battery according to the heat to be controlled.

[0068] In this step, the battery's heat dissipation and temperature equalization structure (such as heat pipes, graphite heat spreaders, etc.) can start a combination of active (such as air cooling) and passive heat dissipation according to the heat to be controlled. The heat pipe conducts core heat, and air cooling accelerates heat dissipation, reducing the temperature difference to ≤ 2°C, ensuring temperature balance during the charging process.

[0069] For example, assuming the heat quantity to be controlled is q = 600W, a start command can be sent to the active cooling module (i.e., the air cooling module) in the heat dissipation and temperature equalization structure. This starts the air cooling module's fan, and the fan speed is adjusted according to the amount of heat to be controlled. In this case, to dissipate 600W of heat, the fan speed is set to 3000 rpm. Simultaneously, the passive cooling components in the heat dissipation structure, such as the heat pipe, thermal grease, and graphite heat spreader, work in tandem. The heat pipe quickly transfers heat from the core area of ​​the battery module to the heat dissipation point. The thermal grease fills the gap between the heat pipe and the battery module, reducing contact thermal resistance. The graphite heat spreader ensures a more uniform temperature distribution across the battery module surface. Through this combination of active and passive cooling, the temperature difference across the battery module gradually decreases. After a period of time, Tcore = 33°C, Tsink = 32.5°C, and ΔT = 0.5°C ≤ 2°C, completing the temperature control process.

[0070] Optionally, the method may further include the following steps: S310: Obtain the actual remaining service life of the battery.

[0071] In this step, the actual remaining service life L real For example, it can be the number of cycles the battery undergoes in actual use until the capacity decays to 20% of the nominal capacity, which is obtained through long-term operation records or full life cycle testing, and the unit is the number of cycles.

[0072] S320. Obtain a model optimization objective function of the remaining useful life prediction model according to the actual remaining useful life, the remaining useful life prediction result, a self-learning weight parameter set, and a regularization term weight factor.

[0073] S321, calculating and obtaining an error between the actual remaining useful life and the remaining useful life prediction result; In this step, the error result is the absolute value of the predicted value and the actual value, that is, |L'-L real |, for example, L'=480 times, L real =500 times, the error result is 20.

[0074] S322. Adjusting the self-learning weight parameter set by the regularization term weight factor to obtain an adjusted self-learning weight parameter set; In this step, the self-learning weight parameter set θ is all the trainable parameters in the model, and the regularization weight factor γ (such as 0.0001) is used to adjust θ. The adjusted set is ,in is the L2 norm of θ (the square root of the sum of the squares of the parameters).

[0075] S323. Based on the error result and the adjusted self-learning weight parameter set, obtain the model optimization objective function of the remaining useful life prediction model.

[0076] In this step, the model optimization objective function L is the sum of the error result and the adjusted self-learning weight parameter set, that is, , which is used to balance prediction accuracy and model complexity to avoid overfitting.

[0077] S330: Optimize the remaining useful life prediction model based on the model optimization objective function.

[0078] In this step, the stochastic gradient descent algorithm can be used to minimize the model optimization objective function L, and the self-learning weight parameter set θ is updated (online fine-tuning every 100 charge and discharge cycles). The prediction accuracy and environmental adaptability of the model are improved through iterative optimization.

[0079] For example, a random gradient descent algorithm can be used to optimize the model, with a learning rate of 0.001 and 100 training cycles. In each training cycle, the training set data is input into the model in batches, each batch containing 32 data, the model optimization objective function L of each batch of data is calculated, and then the self-learning weight parameter set θ of the model is updated according to L, and the update direction is the negative gradient direction of L. After each training cycle, the model is evaluated using the validation set data, and the average error on the validation set is calculated. When the average error of the validation set does not decrease for 10 consecutive training cycles, the training is stopped, and the model obtained at this time is the trained remaining useful life prediction model. In addition, during the actual operation of the battery, every 100 charge and discharge cycles, the adaptive learning module is started, the actual remaining useful life and the prediction result in the 100 cycles are obtained, and the model parameters are fine-tuned online according to the above model optimization objective function to adapt to the characteristic changes of the battery in different use stages.

[0080] The structures described in the above embodiments are connected through signal lines, control lines and power lines to complete data transmission and control logic linkage, realize a closed-loop structure of prediction-feedback-regulation. The method can stably work at an ambient temperature of -40°C to +55°C, and under a 0.5 C charge and discharge rate condition, the cycle life improvement coefficient of the battery is greater than or equal to 1.25 (i.e. the ratio of the cycle life of the battery under the control of the method to the cycle life of the battery under the traditional constant voltage current limiting strategy is greater than or equal to 1.25). The battery service life can be effectively improved, and the replacement frequency and maintenance cost can be reduced under the premise of meeting the backup power supply task.

[0081] The method provided by the embodiments of the present application acquires battery state data in the battery operation process, the battery state data includes voltage state data, current state data, temperature state data and impedance state data of the battery. The state of health SOH of the battery is calculated. The target feature sequence is constructed according to the battery state data and the SOH, and the target feature sequence is input into the pre-trained remaining useful life prediction model to obtain the remaining useful life prediction result of the battery. The charging strategy of the battery is adjusted according to the state of health SOH of the battery and the remaining useful life prediction result, the target charging strategy is obtained, and the battery is charged based on the target charging strategy, so that the charging process always fits the current health state and life expectancy of the battery by dynamically adapting the upper and lower limits of the charging voltage, the current intensity and the charging time, avoiding the overcharging risk caused by fixed parameters in the traditional constant voltage current limiting strategy or the intensified side reaction and thermal runaway hidden danger caused by still using a high-intensity charging mode in the battery aging stage, ensuring that the chemical reaction of the battery is more stable and the structure is more stable in the whole life cycle, reducing unnecessary energy loss and attenuation damage, thereby improving the stability and life of the battery system.

[0082] Figure 2 This is a schematic diagram of the structure of an artificial intelligence-based battery charging control system 100 provided in an embodiment of the present application. Figure 2 As shown, the processor 120 can be used in the artificial intelligence-based battery charging control system 100 and used to perform the functions of the present invention.

[0083] The artificial intelligence-based battery charging control system 100 can be a general-purpose server or a special-purpose server, both of which can be used to implement the artificial intelligence-based battery charging control method of the present invention. Although only one server is shown in the present invention, for convenience, the functions described in the present invention can be implemented in a distributed manner on multiple similar platforms to balance the processing load.

[0084] For example, the artificial intelligence-based battery charging control system 100 may include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and various forms of storage media 140, such as a disk, ROM, or RAM, or any combination thereof. For example, the artificial intelligence-based battery charging control system 100 may also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The method of the present invention may be implemented based on these program instructions. The artificial intelligence-based battery charging control system 100 also includes an input / output (I / O) interface 150 between the computer and other input / output devices.

[0085] For ease of explanation, only one processor is described in the artificial intelligence-based battery charging control system 100. However, it should be noted that the artificial intelligence-based battery charging control system 100 of the present invention may also include multiple processors, so the steps performed by one processor described in the present invention may also be performed jointly or individually by multiple processors. For example, if the processor of the artificial intelligence-based battery charging control system 100 performs steps A and B, it should be understood that steps A and B may also be performed jointly by two different processors or individually in one processor. For example, the first processor performs step A and the second processor performs step B, or the first processor and the second processor perform steps A and B together.

[0086] Optionally, the inner wall of the shell of the artificial intelligence-based battery charging control system 100 can be coated with phase-change thermal conductive silicone grease, the thermal capacity of which is , in order to suppress transient temperature rise.

[0087] Optionally, the artificial intelligence-based battery charging control system 100 can be connected to the communication base station network management system via an RS-485 or Ethernet interface, and supports uploading the remaining service life prediction results and SOH curve using the Modbus-TCP protocol.

[0088] An embodiment of the present invention discloses a computer-readable storage medium storing a computer program for electronic data exchange, wherein the computer program enables a computer to execute the steps of the battery charging control method based on artificial intelligence of the aforementioned embodiment.

[0089] An embodiment of the present invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute the steps in the artificial intelligence-based battery charging control method of the aforementioned embodiment.

[0090] The device embodiments described above are merely illustrative. Modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical modules, i.e., they may be located in one place or distributed across multiple network modules. Some or all of these modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0091] Through the detailed description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus the necessary general hardware platform, or of course, by means of hardware. Based on this understanding, the above technical solution, in essence, or the portion that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM), or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium capable of storing or storing data.

[0092] Finally, it should be noted that what is disclosed above is only a preferred embodiment of the present invention, which is only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A battery charging control method based on artificial intelligence, characterized in that: The method comprises: Acquiring battery status data during battery operation, wherein the battery status data includes voltage status data, current status data, temperature status data, and impedance status data of the battery; Calculating the state of health (SOH) of the battery; Constructing a target feature sequence based on the battery status data and the SOH, and inputting the target feature sequence into a pre-trained remaining service life prediction model to obtain a remaining service life prediction result of the battery; Adjusting the charging strategy of the battery according to the state of health (SOH) of the battery and the remaining service life prediction result to obtain a target charging strategy; The battery is charged based on the target charging strategy.

2. The artificial intelligence-based battery charging control method according to claim 1, characterized in that: Calculating the battery's state of health (SOH) includes: Obtaining the nominal capacity of the battery; The SOH of the battery is calculated based on the battery nominal capacity, the temperature state data, the current state data, and preset parameters, where the preset parameters include a preset weight coefficient, an activation energy constant, and a gas constant.

3. The battery charging control method based on artificial intelligence according to claim 1, characterized in that: The step of constructing a target feature sequence based on the battery status data and the SOH, and inputting the target feature sequence into a pre-trained remaining service life prediction model to obtain a remaining service life prediction result of the battery includes: Extracting local features of the feature sequence through a one-dimensional convolutional neural network in the remaining useful life prediction model; Extracting temporal dependency features in the feature sequence through a bidirectional gated recurrent unit network in the remaining useful life prediction model; Determining feature weights through an attention mechanism in the remaining useful life prediction model; The remaining service life prediction result of the battery is generated according to the local features, the timing dependency features, and the feature weights.

4. The artificial intelligence-based battery charging control method according to claim 1, characterized in that: The step of adjusting the battery charging strategy according to the battery SOH and the remaining service life prediction result to obtain a target charging strategy includes: If the remaining service life prediction result is less than or equal to the preset remaining service life, or the SOH is less than or equal to the preset SOH, obtaining a target charging current according to the SOH, the SOH health threshold, and the SOH change rate; Based on the target charging current, obtaining a target charging output duty cycle; The target charging strategy is determined according to the target charging current and the target charging output duty cycle.

5. The artificial intelligence-based battery charging control method according to claim 4, characterized in that: Obtaining a target charging current according to the SOH, the SOH health threshold, and the SOH change rate includes: Obtaining a first experience parameter and a second experience parameter; Calculating a quotient of the SOH and the SOH health threshold; The target charging current is calculated according to the first empirical parameter, the second empirical parameter, a quotient of the SOH and the SOH health threshold, and the SOH change rate.

6. The artificial intelligence-based battery charging control method according to claim 4, characterized in that: If the remaining useful life prediction result is less than or equal to a preset remaining useful life, or the SOH is less than or equal to a preset SOH, the method further includes: Output charging strategy prompt information, where the charging strategy prompt information is used to prompt the user that the battery will switch to a life-extending slow charging mode.

7. The artificial intelligence-based battery charging control method according to claim 1, characterized in that: After charging the battery based on the target charging strategy, the method further includes: obtaining the core temperature and heat dissipation point temperature of the battery during the charging process; If the difference between the core temperature and the heat dissipation point temperature is greater than or equal to a preset temperature fluctuation threshold, the amount of heat to be controlled is calculated based on the core temperature and the heat dissipation point temperature; Temperature control is performed on the charging process of the battery according to the heat to be controlled.

8. The battery charging control method based on artificial intelligence according to claim 1, characterized in that: The method further comprises: Obtaining the actual remaining service life of the battery; Obtaining a model optimization objective function of the remaining useful life prediction model according to the actual remaining useful life, the remaining useful life prediction result, the self-learning weight parameter set, and the regularization term weight factor; The remaining useful life prediction model is optimized based on the model optimization objective function.

9. The artificial intelligence-based battery charging control method according to claim 8, characterized in that: The method of obtaining a model optimization objective function of the remaining useful life prediction model according to the actual remaining useful life, the remaining useful life prediction result, the self-learning weight parameter set, and the regularization term weight factor includes: Calculating the error between the actual remaining useful life and the remaining useful life prediction result; Adjusting the self-learning weight parameter set by the regularization term weight factor to obtain an adjusted self-learning weight parameter set; Based on the error result and the adjusted self-learning weight parameter set, the model optimization objective function of the remaining useful life prediction model is obtained.

10. A battery charging control system based on artificial intelligence, characterized in that: The invention comprises a processor and a computer-readable storage medium, wherein the computer-readable storage medium stores machine-executable instructions, and when the machine-executable instructions are executed by a computer, the battery charging control method based on artificial intelligence according to any one of claims 1 to 9 is implemented.

Images