Medical simulator battery life prediction method based on dynamic weight and physical constraints

Abstract

The present invention discloses a medical simulator battery life prediction method based on dynamic weights and physical constraints, relating to the field of medical device battery health management. The method includes establishing dynamic fusion weights, a battery capacity physical prediction model, and a battery capacity LSTM prediction model based on the physical characteristics of the medical simulator battery; fusing the battery capacity physical prediction model and the LSTM battery capacity prediction model through the dynamic fusion weights, and establishing a physical constraint loss function, an LSTM loss function, and a joint loss function. The parameters of the prediction model are updated through the loss function, ultimately obtaining a medical simulator battery life prediction result. By fusing the battery capacity physical prediction model and the LSTM battery capacity prediction model, the present invention addresses the problems of insufficient capture of temporal features in traditional LSTM models and the neglect of battery physical mechanisms in pure data-driven methods, thereby improving the accuracy and reliability of medical simulator battery life prediction.

Description

Technical Field

[0001] The present invention relates to the field of medical device battery health management, and in particular to a medical simulator battery life prediction method based on dynamic weights and physical constraints. Background Art

[0002] Medical simulators are devices used to train medical professionals. They simulate real-life medical scenarios, enabling healthcare providers to learn, practice, and evaluate medical skills in a safe environment. This helps limit the risk of accidents for inadequately trained healthcare professionals. Simulator batteries operate over a wide temperature range, typically 10°C to 40°C, and must maintain stability under simulated extreme test conditions. Their charge and discharge cycle is high-frequency, with 3 to 5 charge and discharge cycles per test cycle, significantly different from the daily use of real medical device batteries. This requires support for long-term continuous testing, and battery life directly impacts simulator testing accuracy. Accurately predicting the life of medical simulator batteries helps plan battery maintenance and replacement in advance, reducing equipment downtime and ensuring smooth medical training and testing processes. This also improves the safety and performance of medical simulators.

[0003] Existing technologies for predicting medical simulator battery life have numerous shortcomings. Traditional threshold methods are overly simplistic, only able to identify significant drops in battery capacity and failing to capture the subtle capacity decay of simulator batteries. Industrial hybrid models are insufficiently capable of capturing the high-frequency charging and discharging of medical simulators, resulting in high false positive rates and low prediction accuracy. Purely physical models, such as the Arrhenius equation, are based on theoretical assumptions and ignore the nonlinear characteristics of actual medical simulator test data, limiting prediction accuracy. Furthermore, traditional LSTM models fail to adequately capture the timing characteristics of high-frequency charging and discharging in medical simulators, failing to accurately match the actual operating conditions of simulator batteries. Furthermore, purely data-driven methods ignore the nonlinear effects of temperature gradients on the thickening rate of the battery's SEI film, weakening the accuracy and reliability of predictions. Summary of the Invention

[0004] The present invention provides a medical simulator battery life prediction method based on dynamic weights and physical constraints. By integrating the battery capacity physical prediction model and the LSTM battery capacity prediction model, the present invention solves the problems of insufficient capture of timing features of the traditional LSTM model and the pure data-driven method ignoring the physical mechanism of the battery, thereby improving the accuracy and reliability of the medical simulator battery life prediction.

[0005] To achieve the above objectives, the present invention provides a medical simulator battery life prediction method based on dynamic weights and physical constraints, comprising:

[0006] According to the physical characteristics of the medical simulator battery, dynamic fusion weights, battery capacity physical prediction model and battery capacity LSTM prediction model are established;

[0007] According to the dynamic fusion weight, the physical prediction capacity of the battery capacity physical prediction model and the LSTM prediction capacity of the battery capacity LSTM prediction model are fused to obtain the final battery capacity prediction result;

[0008] Based on the final battery capacity prediction results, a physical constraint loss function of the battery capacity physical prediction model and an LSTM loss function of the battery capacity LSTM prediction model are established. A joint loss function is established based on the physical constraint loss function and the LSTM loss function.

[0009] The parameters of the battery capacity physical prediction model are updated according to the joint loss function, and the parameters of the battery capacity LSTM prediction model are updated according to the joint loss function to obtain an updated battery capacity physical prediction model and an updated battery capacity LSTM prediction model;

[0010] The updated battery capacity physical prediction model and the updated battery capacity LSTM prediction model are used to process the medical simulator battery to obtain the medical simulator battery life prediction results.

[0011] In view of the characteristics of medical simulator batteries, such as wide operating temperature range, high-frequency intermittent charging and discharging, and long continuous working time, the present invention proposes a medical simulator battery life prediction method based on dynamic weights and physical constraints. The battery life of the medical simulator is predicted by fusing a battery capacity physical prediction model and an LSTM battery capacity prediction model. The battery capacity physical prediction model can predict the battery capacity of the medical simulator within extreme operating temperatures, so that the battery physical predicted capacity is consistent with the actual aging rate of the medical simulator at different temperatures. The LSTM battery capacity prediction model can capture the timing characteristics of battery capacity changes and obtain accurate battery LSTM predicted capacity. Based on the high-frequency intermittent charging and discharging unique to the medical simulator, dynamic fusion weights related to the medical simulator battery charging and discharging interval and average current are designed. The battery physical predicted capacity and the battery LSTM predicted capacity are fused through the dynamic fusion weights, thereby reducing the battery capacity prediction error of the medical simulator under high-frequency charging and discharging and long working time, and making the prediction results consistent with the battery physical and chemical change process within the medical simulator working temperature, thereby improving the accuracy and reliability of the medical simulator battery life prediction.

[0012] Furthermore: the expression of the dynamic fusion weight is as follows:

[0013] ;

[0014] in, is the dynamic fusion weight, is the Sigmoid function, is the weight matrix of dynamic fusion weights, is the bias term of dynamic fusion weight, is the average current of the medical simulator battery, is the charge and discharge interval of the medical simulator battery, is a vector consisting of the average current and charge and discharge intervals of the medical simulator battery.

[0015] The present invention uses the average current and charge and discharge interval of the medical simulator battery as input vectors, and uses the Sigmoid function, weight matrix and bias term to calculate the dynamic fusion weight, which solves the problem in the existing technology that the traditional LSTM model does not adequately capture the high-frequency charge and discharge timing characteristics of the medical simulator and does not consider the battery attenuation rate, resulting in inaccurate battery life prediction results; the dynamic fusion weight can be dynamically adjusted according to the different conditions of the medical simulator battery, dynamically balancing the contribution of the battery capacity physical prediction model and the battery capacity LSTM prediction model, improving the model's adaptability to high-frequency intermittent charge and discharge scenarios, and enhancing the accuracy of battery life prediction.

[0016] Furthermore: the expression of the battery capacity physical prediction model is as follows:

[0017] ;

[0018] ;

[0019] ;

[0020] in, Battery for medical simulator Physical prediction capacity of the first charge and discharge cycle, is the initial capacity of the medical simulator battery, is a natural constant, is the number of charge and discharge cycles of the medical simulator, is the temperature-dependent aging rate, is the frequency factor, For the Activation energy of medical simulator battery after 1 charge-discharge cycle, For the Temperature variation compensation coefficient for each charge and discharge cycle, is the temperature fluctuation amplitude of a single charging cycle, is the gas constant, is the battery temperature, is the maximum temperature of a single charging cycle, The minimum temperature for a single charging cycle.

[0021] The present invention modifies the traditional Arrhenius equation by introducing a temperature fluctuation compensation coefficient, and correlates battery capacity decay with battery temperature, activation energy and temperature fluctuation amplitude, reflecting the aging rate of medical simulator batteries at different temperatures, and accurately capturing the impact of temperature changes on battery capacity decay. This solves the problem that traditional methods cannot accurately describe the battery capacity decay characteristics of medical simulator batteries under complex temperature environments and high-frequency charging and discharging conditions, improves the prediction accuracy of medical simulator batteries under extreme operating temperature conditions, and makes the battery capacity prediction results more in line with the battery aging laws in actual usage scenarios.

[0022] Furthermore: the expression of the battery capacity LSTM prediction model is as follows:

[0023] ;

[0024] ;

[0025] ;

[0026] ;

[0027] in, Battery for medical simulator LSTM predicted capacity of charge and discharge cycles, are trainable weights, is the bias, For the LSTM hidden state of the charge-discharge cycle, For the LSTM hidden state of the charge-discharge cycle, is the LSTM network model function, is the capacity sequence, is the temperature, is the charge and discharge current, To reset the gate, To reset the gate weights, To reset the gate bias, is the Sigmoid activation function, For candidate gates, is the weight of the candidate gate, is the bias of the candidate gate, is the hyperbolic tangent function.

[0028] The battery capacity LSTM prediction model of the present invention uses data such as capacity sequence, temperature, and charge and discharge current, and utilizes the gating mechanism of the LSTM network to capture timing dependencies, and learns the complex dynamic characteristics of the medical simulator battery during high-frequency intermittent charging and discharging. It can solve the problem that pure data-driven methods cannot fully utilize the physical and chemical mechanism information of the battery. At the same time, it overcomes the defect of the existing technology in insufficiently capturing the high-frequency charging and discharging timing characteristics when processing the specific working conditions of the medical simulator, improves the prediction accuracy of the battery capacity change trend, and adapts to the actual use conditions of the medical simulator battery under high-frequency intermittent charging and discharging scenarios.

[0029] Furthermore, the final battery capacity prediction result is expressed as follows:

[0030] ;

[0031] in, Battery for medical simulator The final predicted battery capacity after two charge and discharge cycles, is the dynamic fusion weight, Battery for medical simulator Physical prediction capacity of the charge-discharge cycle, Battery for medical simulator LSTM predicted capacity of the charge-discharge cycle.

[0032] The expression of the final battery capacity prediction result of the present invention performs weighted summation of the physical model predicted capacity and the LSTM predicted capacity through dynamic fusion weights, thereby realizing the combination of the battery capacity physical prediction model and the battery capacity LSTM prediction model. By utilizing the in-depth analysis of the battery aging mechanism by the battery capacity physical prediction model and the fitting ability of the battery capacity LSTM prediction model to complex battery time series data, the problem that traditional methods are difficult to balance the battery physical laws and data characteristics under different medical simulator working conditions is solved. By dynamically adjusting the weights, the medical simulator can better adapt to actual usage scenarios under complex working conditions such as high-frequency intermittent charging and discharging conditions and extreme operating temperatures, thereby improving the accuracy and reliability of battery capacity prediction.

[0033] Furthermore, the expression of the physical constraint loss function is as follows:

[0034] ;

[0035] ;

[0036] ;

[0037] ;

[0038] in, is the physical constraint loss function, is the first weight parameter, To predict the battery capacity attenuation gradient, is the actual battery capacity attenuation gradient.

[0039] In response to the problem of insufficient prediction accuracy caused by the pure data-driven method in the prior art when the physical and chemical processes of the battery are ignored, the present invention designs a physical constraint loss function for the battery capacity physical prediction model. By introducing the physical constraint term, the difference between the predicted battery capacity attenuation gradient and the actual battery capacity attenuation gradient is included in the calculation, so that the modified Arrhenius equation fits the actual battery aging rate under extreme temperatures, thereby improving the prediction accuracy of the battery capacity physical prediction model and enhancing the reliability of the technical solution of the present invention under complex working conditions. This solves the problem that traditional methods cannot effectively utilize the physical laws of batteries to constrain model training to improve prediction accuracy.

[0040] Furthermore: the expression of the LSTM loss function is as follows:

[0041] ;

[0042] ;

[0043] in, is the LSTM loss function, is the gradient of each LSTM gating unit, are LSTM network parameters, is the second weight parameter, is the final battery capacity prediction result, is the actual battery capacity.

[0044] The present invention addresses the problem in the prior art that traditional LSTM models are difficult to adapt to the complex high-frequency charging and discharging conditions of medical simulators. By using the LSTM loss function, the mean square error between the final battery capacity prediction result and the actual battery capacity is quantified, and the gradient of the error relative to the LSTM network parameters is calculated. This solves the shortcomings of traditional methods in capturing timing features, promotes the adjustment of LSTM network parameters, enables the battery capacity LSTM prediction model to better learn the timing features of battery capacity changes, and improves the prediction accuracy and adaptability of the battery capacity LSTM prediction model under high-frequency intermittent charging and discharging conditions.

[0045] Furthermore, the expression of the joint loss function is as follows:

[0046] ;

[0047] in, is the joint loss function, is the third weight parameter, For the LSTM predicted battery capacity, is the actual battery capacity, is the predicted battery capacity attenuation gradient, is the actual battery capacity attenuation gradient.

[0048] The present invention combines the physical prediction of battery capacity with the LSTM prediction of battery capacity through a joint loss function, while taking into account the deviation between the predicted battery capacity and the actual battery capacity and the difference between the predicted capacity attenuation gradient and the actual capacity attenuation gradient, thereby solving the problem of inaccurate prediction of a single model in the prior art and the problem that the traditional method is difficult to balance the battery data characteristics and the physical change laws of the battery; through the joint loss function, the parameters of the physical prediction model of battery capacity and the parameters of the LSTM prediction model of battery capacity can be optimized at the same time, while adapting to the complex working conditions of high-frequency intermittent charging and discharging, ensuring that the prediction results conform to the actual aging laws of the battery, thereby improving the accuracy and reliability of battery life prediction.

[0049] Furthermore, the updating of the parameters of the battery capacity physical prediction model according to the joint loss function specifically includes:

[0050] The temperature change compensation coefficient of the battery capacity physical prediction model is updated. The expression is as follows:

[0051] ;

[0052] in, For the Temperature variation compensation coefficient for each charge and discharge cycle, For the Temperature variation compensation coefficient for each charge and discharge cycle, is the learning rate of the temperature change compensation coefficient, is the joint loss function;

[0053] The activation energy of the battery capacity physical prediction model is updated as follows:

[0054] ;

[0055] in, For the The activation energy of the charge-discharge cycle after updating based on the physical constraint loss function, For the The activation energy of the charge-discharge cycle after updating based on the physical constraint loss function, is the learning rate of the activation energy.

[0056] The present invention uses a gradient descent method and a joint loss function to dynamically adjust the temperature change compensation coefficient and activation energy of the battery capacity physical prediction model, thereby solving the problems in the prior art that the physical model parameters cannot adapt to the actual operating conditions of the medical simulator battery, and the problem that the physical model parameters are difficult to correct under different temperatures and charge and discharge modes; the present invention updates the temperature change compensation coefficient and activation energy, so that the battery capacity physical prediction model can better reflect the aging characteristics of the actual battery, and has higher prediction accuracy and adaptability under extreme operating temperatures and high-frequency charge and discharge conditions.

[0057] Furthermore, the parameters of the battery capacity LSTM prediction model are updated according to the joint loss function, and the expression is as follows:

[0058] ;

[0059] ;

[0060] ;

[0061] in, is the updated LSTM network parameter, is the LSTM network parameter before updating, is the learning rate of the battery capacity LSTM prediction model, For the The first-order moment of the gradient of the charge-discharge cycle, For the The first-order moment of the charge-discharge cycle, For the The second-order moment of the charge-discharge cycle, For the The second-order moment of the charge-discharge cycle, is a constant, is the first-order moment decay rate, is the second-order moment decay rate, For the The gradient of the LSTM network parameters for each charge and discharge cycle, For the Element-wise squared gradient of the LSTM network parameters for charge-discharge cycles.

[0062] The present invention updates the parameters of the battery capacity LSTM prediction model based on a joint loss function and an adaptive learning rate method of the Adam optimizer. The LSTM network parameters are adjusted by estimating the first-order and second-order moments of the gradient, so that the LSTM model can simultaneously consider data fitting and physical law constraints, thereby improving the prediction accuracy of the battery capacity LSTM prediction model for battery capacity change trends. This solves the problem in the prior art that traditional LSTM models have difficulty capturing timing characteristics and adapting to high-noise data when processing the complex working conditions of high-frequency intermittent charging and discharging of medical simulator batteries, resulting in insufficient prediction accuracy.

[0063] One or more technical solutions provided by the present invention have at least the following technical effects or advantages:

[0064] Based on the charge and discharge intervals and average current of the medical simulator battery, the present invention designs dynamic fusion weights, and fuses the battery capacity physical prediction model with the battery capacity LSTM prediction model through the dynamic fusion weights to construct a hybrid battery life prediction model architecture. The dynamic weight adjustment mechanism and the joint loss function are introduced to enable the hybrid battery life prediction model architecture to adapt to the high-frequency intermittent charge and discharge characteristics of the medical simulator battery, accurately capture the timing characteristics, and at the same time combine physical constraints to improve the prediction accuracy and reliability, so that the battery capacity prediction results fit the battery aging law of the medical simulator battery in the actual usage scenario, thereby improving the accuracy and reliability of the medical simulator battery life prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of the present invention, and do not constitute a limitation of the embodiments of the present invention;

[0066] Figure 1 This is a flow chart of the medical simulator battery life prediction method based on dynamic weights and physical constraints in the present invention. DETAILED DESCRIPTION

[0067] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present invention and the features therein can be combined with each other without conflict.

[0068] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present invention is not limited to the specific embodiments disclosed below.

[0069] Example 1

[0070] like Figure 1As shown, the present invention provides a medical simulator battery life prediction method based on dynamic weights and physical constraints, including:

[0071] According to the physical characteristics of the medical simulator battery, dynamic fusion weights, battery capacity physical prediction model and battery capacity LSTM prediction model are established;

[0072] According to the dynamic fusion weight, the physical prediction capacity of the battery capacity physical prediction model and the LSTM prediction capacity of the battery capacity LSTM prediction model are fused to obtain the final battery capacity prediction result;

[0073] Based on the final battery capacity prediction results, a physical constraint loss function of the battery capacity physical prediction model and an LSTM loss function of the battery capacity LSTM prediction model are established. A joint loss function is established based on the physical constraint loss function and the LSTM loss function.

[0074] The parameters of the battery capacity physical prediction model are updated according to the joint loss function, and the parameters of the battery capacity LSTM prediction model are updated according to the joint loss function to obtain an updated battery capacity physical prediction model and an updated battery capacity LSTM prediction model;

[0075] The updated battery capacity physical prediction model and the updated battery capacity LSTM prediction model are used to process the medical simulator battery to obtain the medical simulator battery life prediction results.

[0076] In daily use, the battery characteristics of medical simulators include: a wide operating temperature range, capable of operating in an environment of -10℃~40℃; the charging and discharging mode is high-frequency intermittent, requiring 3 to 5 charging and discharging times within the medical simulator test cycle, which is significantly different from the daily use of real medical equipment; long-term continuous testing requires the medical simulator to run uninterruptedly for 72 hours; the life of the medical simulator battery will affect the test accuracy of the medical simulator.

[0077] Aiming at the battery characteristics of medical simulators, this paper proposes a medical simulator battery life prediction method based on dynamic weights and physical constraints. By integrating the battery capacity physical prediction model and the LSTM battery capacity prediction model, a hybrid model architecture is constructed to predict the battery life of the medical simulator. The battery capacity physical prediction model is based on the Arrhenius equation and introduces a temperature fluctuation compensation coefficient for correction. The battery capacity physical prediction model is established, which can accurately reflect the aging and attenuation law of the battery in the temperature range of -10℃~40℃, and adapt to the extreme working temperature requirements of the medical simulator; the battery capacity LSTM prediction model can capture the timing characteristics of high-frequency intermittent charging and discharging, solve the problem of insufficient timing feature capture of the traditional LSTM model, and adapt to the special charging and discharging mode of the medical simulator; introduce a dynamic fusion weight adjustment mechanism, and adaptively adjust the dynamic fusion weight according to the charging and discharging interval and average current, so that the model dynamically balances the contributions of the battery capacity physical prediction model and the battery capacity LSTM prediction model under different working conditions, enhances the prediction performance under high-frequency charging and discharging conditions, and improves the model adaptability; through the joint loss function, the parameters of the battery capacity physical prediction model and the battery capacity LSTM prediction model are synchronously updated to ensure that the prediction results fit the actual working conditions of the battery and improve the accuracy and reliability of battery life prediction.

[0078] In the present invention, the charge and discharge intervals of the medical simulator test cycle are and average current Get the dynamic fusion weight, the expression of dynamic fusion weight is as follows:

[0079] ;

[0080] in, is the dynamic fusion weight, is the Sigmoid function, is the weight matrix of dynamic fusion weights, is the bias term of dynamic fusion weight, is the average current of the medical simulator battery, is the charge and discharge interval of the medical simulator battery, It is a vector composed of the average current and charge and discharge interval of the medical simulator battery. When the medical simulator is in high-frequency mode, the charge and discharge interval of the medical simulator battery is generally less than 10 minutes. At this time, the dynamic fusion weight is about 0.1, which can enhance the battery capacity LSTM prediction model's ability to capture instantaneous current pulses. When the medical simulator is in normal mode, the charge and discharge interval of the medical simulator battery is generally greater than 30 minutes. At this time, the dynamic fusion weight is about 0.5, which can balance the contributions of the battery capacity physical prediction model and the battery capacity LSTM prediction model. The dynamic fusion weight of the present invention can be dynamically adjusted according to the different conditions of the medical simulator battery, improving the adaptability of the battery capacity physical prediction model and the battery capacity LSTM prediction model to high-frequency intermittent charge and discharge scenarios, and improving the accuracy of battery life prediction, solving the problems in the prior art where the traditional LSTM model does not capture the high-frequency charge and discharge timing characteristics of the medical simulator and does not consider the battery attenuation rate.

[0081] In the present invention, a temperature fluctuation compensation coefficient is introduced to correct the Arrhenius equation to obtain a battery capacity physical prediction model. This model can associate battery capacity decay with battery temperature, activation energy, and temperature fluctuation amplitude, indirectly reflecting the aging rate of medical simulator batteries at different temperatures and capturing the impact of temperature changes on medical simulator battery capacity decay. This solves the problem that traditional methods cannot accurately describe the battery capacity decay characteristics of medical simulator batteries under complex temperature environments and high-frequency charging and discharging conditions, making the battery capacity prediction results more consistent with the battery aging laws in actual usage scenarios. The expression of the battery capacity physical prediction model of the present invention is as follows:

[0082] ;

[0083] ;

[0084] ;

[0085] in, Battery for medical simulator The physical predicted capacity of the first charge and discharge cycle, in mAh, is the initial capacity of the medical simulator battery, in mAh. is a natural constant, is the number of charge and discharge cycles of the medical simulator, that is, the number of test cycles of the medical simulator, is the temperature-dependent aging rate in / s, is the frequency factor, For the The activation energy of the medical simulator battery after 1 charge and discharge cycle represents the degree of material aging of the medical simulator battery. For the The temperature change compensation coefficient of the charge and discharge cycle can be set to 0.67 based on experience. is the temperature fluctuation amplitude of a single charging cycle, in K, is the gas constant, is the battery temperature in K, is the maximum temperature of a single charging cycle, in K, The minimum temperature during a single charge cycle, in K.

[0086] In the present invention, the capacity sequence ,temperature and charge and discharge current The battery LSTM capacity is used as the input of the LSTM network and as the output of the LSTM network. The gating mechanism of the LSTM network is used to capture the timing dependency of the medical simulator battery and learn the dynamic characteristics of the medical simulator battery during high-frequency intermittent charging and discharging. This solves the problem that existing technologies cannot capture the high-frequency charging and discharging timing characteristics when dealing with the specific working conditions of the medical simulator. It improves the prediction accuracy of the battery capacity change trend and adapts to the actual use conditions of the medical simulator battery under high-frequency intermittent charging and discharging scenarios. The expression of the battery capacity LSTM prediction model is as follows:

[0087] ;

[0088] ;

[0089] ;

[0090] ;

[0091] in, Battery for medical simulator LSTM predicted capacity of charge and discharge cycles, are trainable weights, is the bias, For the LSTM hidden state of the charge-discharge cycle, For the LSTM hidden state of the charge-discharge cycle, is the LSTM network model function, is the capacity sequence, is the temperature, is the charge and discharge current, To reset the gate, To reset the gate weights, To reset the gate bias, is the Sigmoid activation function, For candidate gates, is the weight of the candidate gate, is the bias of the candidate gate, is the hyperbolic tangent function.

[0092] According to the dynamic fusion weight, the physical prediction capacity of the battery capacity physical prediction model and the LSTM prediction capacity of the battery capacity LSTM prediction model are fused to obtain the final battery capacity prediction result. The expression of the final battery capacity prediction result is as follows:

[0093] ;

[0094] in, Battery for medical simulator The final predicted battery capacity after two charge and discharge cycles, is the dynamic fusion weight, Battery for medical simulator Physical prediction capacity of the charge-discharge cycle, Battery for medical simulator The LSTM prediction capacity of the battery capacity cycle is realized through dynamic fusion weights. The battery capacity physical prediction model and the battery capacity LSTM prediction model are integrated. By using the in-depth analysis of the battery aging mechanism by the battery capacity physical prediction model and the fitting ability of the battery capacity LSTM prediction model to the complex time series data of the battery, the medical simulator can better adapt to the actual use scenarios under complex working conditions such as high-frequency intermittent charging and discharging conditions and extreme working temperatures, thereby improving the accuracy and reliability of battery capacity prediction. Among them, the dynamic fusion weights Represents the contribution of the battery capacity physical prediction model, when the dynamic fusion weight , it means that the final battery predicted capacity is completely dependent on the battery capacity physical prediction model.

[0095] In the present invention, a physical constraint loss function of the battery capacity physical prediction model is established based on back propagation, and the objects of action are the temperature change compensation coefficient and activation energy. The expression of the physical constraint loss function is as follows:

[0096] ;

[0097] ;

[0098] ;

[0099] ;

[0100] in, is the physical constraint loss function, is the first weight parameter, which can be set according to experience. To predict the battery capacity attenuation gradient, is the actual battery capacity attenuation gradient.

[0101] In the present invention, the LSTM loss function of the battery capacity LSTM prediction model is obtained by back propagation, which is used to update the LSTM network parameters, so that the battery capacity LSTM prediction model can better learn the time series characteristics of battery capacity changes, and improve the prediction accuracy and adaptability of the battery capacity LSTM prediction model under high-frequency intermittent charging and discharging conditions. The expression of the LSTM loss function is as follows:

[0102] ;

[0103] ;

[0104] in, is the LSTM loss function, is the gradient of each LSTM gate unit, calculated by time back propagation, involving the LSTM cell state and hidden state The temporal gradient accumulation, are LSTM network parameters, is the second weight parameter, which can be set based on experience. is the final battery capacity prediction result, is the actual battery capacity.

[0105] In the present invention, a joint loss function of the battery capacity physical prediction model and the battery capacity LSTM prediction model is established based on the physical constraint loss function and the LSTM loss function, combining the battery capacity physical prediction with the battery capacity LSTM prediction, and taking into account the deviation between the predicted battery capacity and the actual battery capacity and the difference between the predicted capacity attenuation gradient and the actual capacity attenuation gradient, to solve the problem of inaccurate prediction of a single model in the prior art, and the problem that traditional methods are difficult to balance battery data characteristics and battery physical change laws; in complex working conditions of high-frequency intermittent charging and discharging, it ensures that the prediction results conform to the actual aging laws of the battery, thereby improving the accuracy and reliability of battery life prediction. The expression of the joint loss function of the present invention is as follows:

[0106] ;

[0107] in, is the joint loss function, is the third weight parameter, which can be set according to experience. For the LSTM predicted battery capacity, is the actual battery capacity, is the predicted battery capacity attenuation gradient, is the actual battery capacity attenuation gradient.

[0108] After obtaining the joint loss function, the Adam optimizer is used to update the temperature change compensation coefficient and activation energy based on the joint loss function, so that the battery capacity physical prediction model can better reflect the actual battery aging characteristics. It has higher prediction accuracy and adaptability under extreme operating temperatures and high-frequency charging and discharging conditions, and makes the battery physical predicted capacity fit the actual battery aging rate under extreme temperatures. Specifically, the following are included:

[0109] The temperature change compensation coefficient of the battery capacity physical prediction model is updated. The expression is as follows:

[0110] ;

[0111] in, For the Temperature variation compensation coefficient for each charge and discharge cycle, For the Temperature variation compensation coefficient for each charge and discharge cycle, is the learning rate of the temperature change compensation coefficient, is the joint loss function;

[0112] The activation energy of the battery capacity physical prediction model is updated as follows:

[0113] ;

[0114] in, For the The activation energy of the charge-discharge cycle after updating based on the physical constraint loss function, For the The activation energy of the charge-discharge cycle after updating based on the physical constraint loss function, is the learning rate of the activation energy.

[0115] After obtaining the joint loss function, the parameters of the battery capacity LSTM prediction model are updated using the Adam optimizer's adaptive learning rate method based on the joint loss function. The LSTM network parameters are adjusted by estimating the first-order and second-order moments of the gradient of the joint loss function. This allows the LSTM model to simultaneously consider data fitting and physical constraints, improving the accuracy of the battery capacity LSTM prediction model for predicting battery capacity trends. This addresses the existing problem of traditional LSTM models having difficulty capturing timing characteristics and adapting to high-noise data when processing the complex operating conditions of high-frequency intermittent charging and discharging of medical simulator batteries, resulting in insufficient prediction accuracy. The parameters of the battery capacity LSTM prediction model are updated based on the joint loss function, and the expression is as follows:

[0116] ;

[0117] ;

[0118] ;

[0119] in, is the updated LSTM network parameter, is the LSTM network parameter before updating, is the learning rate of the battery capacity LSTM prediction model, For the The first-order moment of the gradient of the charge-discharge cycle, For the The first-order moment of the charge-discharge cycle, For the The second-order moment of the charge-discharge cycle, For the The second-order moment of the charge-discharge cycle, is a constant, which is usually set to a very small constant based on experience to ensure numerical stability. is the first-order moment decay rate, is the second-order moment decay rate, For the The gradient of the LSTM network parameters for each charge and discharge cycle, For the Element-wise squared gradient of the LSTM network parameters for charge-discharge cycles.

[0120] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0121] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A medical simulator battery life prediction method based on dynamic weights and physical constraints, characterized by: include: According to the physical characteristics of the medical simulator battery, dynamic fusion weights, battery capacity physical prediction model and battery capacity LSTM prediction model are established; According to the dynamic fusion weight, the physical prediction capacity of the battery capacity physical prediction model and the LSTM prediction capacity of the battery capacity LSTM prediction model are fused to obtain the final battery capacity prediction result; Based on the final battery capacity prediction results, a physical constraint loss function of the battery capacity physical prediction model and an LSTM loss function of the battery capacity LSTM prediction model are established. A joint loss function is established based on the physical constraint loss function and the LSTM loss function. The parameters of the battery capacity physical prediction model are updated according to the joint loss function, and the parameters of the battery capacity LSTM prediction model are updated according to the joint loss function to obtain an updated battery capacity physical prediction model and an updated battery capacity LSTM prediction model; The updated battery capacity physical prediction model and the updated battery capacity LSTM prediction model are used to process the medical simulator battery to obtain the medical simulator battery life prediction results; The expression of the dynamic fusion weight is as follows: in, is the dynamic fusion weight, is the Sigmoid function, is the weight matrix of dynamic fusion weights, is the bias term of dynamic fusion weight, is the average current of the medical simulator battery, is the charge and discharge interval of the medical simulator battery, is a vector consisting of the average current and charge and discharge interval of the medical simulator battery; The expression of the joint loss function is as follows: in, is the joint loss function, is the third weight parameter, is the final battery capacity prediction result, is the actual battery capacity, is the predicted battery capacity attenuation gradient, is the actual battery capacity attenuation gradient.

2. The medical simulator battery life prediction method based on dynamic weights and physical constraints according to claim 1 is characterized in that: The expression of the battery capacity physical prediction model is as follows: in, Battery for medical simulator Physical prediction capacity of the first charge and discharge cycle, is the initial capacity of the medical simulator battery, is a natural constant, is the number of charge and discharge cycles of the medical simulator, is the temperature-dependent aging rate, is the frequency factor, For the Activation energy of medical simulator battery after 1 charge-discharge cycle, For the Temperature variation compensation coefficient for each charge and discharge cycle, is the temperature fluctuation amplitude of a single charging cycle, is the gas constant, is the battery temperature, is the maximum temperature of a single charging cycle, The minimum temperature for a single charging cycle.

3. The medical simulator battery life prediction method based on dynamic weights and physical constraints according to claim 1 is characterized in that: The expression of the battery capacity LSTM prediction model is as follows: in, Battery for medical simulator LSTM predicted capacity of charge and discharge cycles, are trainable weights, is the bias, For the LSTM hidden state of the charge-discharge cycle, For the LSTM hidden state of the charge-discharge cycle, is the capacity sequence, is the temperature, is the charge and discharge current, To reset the gate, To reset the gate weights, To reset the gate bias, is the Sigmoid activation function, For candidate gates, is the weight of the candidate gate, is the bias of the candidate gate, is the hyperbolic tangent function.

4. The medical simulator battery life prediction method based on dynamic weights and physical constraints according to claim 1, characterized in that: The final battery capacity prediction result is expressed as follows: in, Battery for medical simulator The final predicted battery capacity after two charge and discharge cycles, is the dynamic fusion weight, Battery for medical simulator Physical prediction capacity of the first charge and discharge cycle, Battery for medical simulator LSTM predicted capacity of the charge-discharge cycle.

5. The medical simulator battery life prediction method based on dynamic weights and physical constraints according to claim 2, characterized in that: The expression of the physical constraint loss function is as follows: in, is the physical constraint loss function, is the first weight parameter, Physically predict capacity of medical simulator batteries, Battery for medical simulator The actual capacity of the charge and discharge cycle.

6. The medical simulator battery life prediction method based on dynamic weights and physical constraints according to claim 3 is characterized in that: The expression of the LSTM loss function is as follows: in, is the LSTM loss function, is the gradient of each LSTM gating unit, are LSTM network parameters, is the second weight parameter.

7. The medical simulator battery life prediction method based on dynamic weights and physical constraints according to claim 1, characterized in that: The updating of the parameters of the battery capacity physical prediction model according to the joint loss function specifically includes: The temperature change compensation coefficient of the battery capacity physical prediction model is updated. The expression is as follows: in, For the Temperature variation compensation coefficient for each charge and discharge cycle, For the Temperature variation compensation coefficient for each charge and discharge cycle, is the learning rate of the temperature change compensation coefficient, is the joint loss function; The activation energy of the battery capacity physical prediction model is updated as follows: in, For the The activation energy of the charge-discharge cycle after updating based on the physical constraint loss function, For the The activation energy of the charge-discharge cycle after updating based on the physical constraint loss function, is the learning rate of the activation energy.

8. The medical simulator battery life prediction method based on dynamic weights and physical constraints according to claim 1 is characterized in that: The parameters of the battery capacity LSTM prediction model are updated according to the joint loss function, and the expression is as follows: in, is the updated LSTM network parameter, is the LSTM network parameter before updating, is the learning rate of the battery capacity LSTM prediction model, For the The first-order moment of the gradient of the charge-discharge cycle, For the The first-order moment of the charge-discharge cycle, For the The second-order moment of the charge-discharge cycle, For the The second-order moment of the charge-discharge cycle, is a constant, is the first-order moment decay rate, is the second-order moment decay rate, For the The gradient of the LSTM network parameters for each charge and discharge cycle, For the Element-wise squared gradient of the LSTM network parameters for charge-discharge cycles.

Images