Safety state diagnosis method and system of lithium iron phosphate battery energy storage system

Abstract

The invention belongs to the technical field of battery safety management, and provides a safety state diagnosis method and system for a lithium iron phosphate battery energy storage system, and the method comprises the steps: obtaining the safety state of a battery according to obtained related parameters and a preset safety state prediction model; during neural network learning, the weight of the current layer is updated according to an error function, the output value of each layer is used as an input value, reverse calculation is carried out by using the updated weight, and the calculated value is used for calculating the error of the previous layer and updating the weight value again; and introducing the correlation between related parameters and the safety state of the battery to correct the updated weight value. On the basis of comprehensively considering voltage, current, temperature, deformation, noise, gas content and dQ / dV-U peak area parameters in a voltage preset interval, considering different influence degrees of different parameter changes on battery cell thermal runaway, and introducing correlation between related parameters and a battery safety state to correct an updated weight value; and the accuracy of a prediction result is ensured.

Description

Technical Field

[0001] The present invention belongs to the technical field of battery safety management, and in particular relates to a safety status diagnosis method and system for a lithium iron phosphate battery energy storage system. Background Art

[0002] The scale of development and utilization of new energy sources such as photovoltaics and wind power continues to expand, and the power storage industry has also developed rapidly. As an important supporting technology for building a new power system, electrochemical energy storage has ushered in an unprecedented period of development opportunities. It is the new energy storage technology with the largest application scale and the fastest growth in recent years. Among them, lithium iron phosphate batteries are the mainstream technical route for electrochemical energy storage due to their advantages such as long discharge time, fast response speed, high conversion efficiency, no restrictions on natural conditions, and easy large-scale application. The installed capacity accounts for a high proportion, but its safety operation problems are gradually highlighted, seriously affecting the safe and stable operation of energy storage power stations, causing serious economic losses and social impacts. Therefore, with the development of the lithium battery energy storage industry, it is of great significance to carry out research on the evaluation and early warning of the safety status of energy storage systems and the control of accident risks. Energy storage systems generally use lithium iron phosphate battery modules. The thermal runaway of individual batteries can easily affect other nearby batteries, eventually causing the entire system to fail to operate normally, and even causing serious safety problems.

[0003] The inventors have found that the main technical route in the current battery energy storage system safety solution is to obtain electrical data such as battery voltage, current and temperature based on the Battery Management System (BMS), and calculate the safety status of the battery cell through a data-driven model. However, its disadvantage is that it does not consider the influence of parameters such as deformation, noise, gas content, and dQ / dV-U peak area within a preset voltage range, and cannot accurately predict the safety of the lithium iron phosphate battery energy storage system. In addition, the correlation between the electrical data such as battery voltage, current and temperature and the battery safety status is not high enough, and the prediction results are often inaccurate. Summary of the invention

[0004] In order to solve the above problems, the present invention proposes a safety status diagnosis method and system for a lithium iron phosphate battery energy storage system. The present invention comprehensively considers voltage, current, temperature, deformation, noise, gas content and dQ / dV-U peak area parameters within a preset voltage range, considers that different parameter changes have different degrees of influence on thermal runaway of the battery cell, introduces the correlation between relevant parameters and the battery safety status to correct the updated weight value, improves the correlation between the parameters and the battery safety status, and ensures the accuracy of the prediction result.

[0005] In order to achieve the above object, the present invention is implemented through the following technical solutions:

[0006] In a first aspect, the present invention provides a safety status diagnosis method for a lithium iron phosphate battery energy storage system, comprising:

[0007] Obtain relevant parameters of the battery, including voltage, current, temperature, deformation, noise, gas content, and dQ / dV-U peak area within a preset voltage range;

[0008] The safety status of the battery is obtained based on the obtained relevant parameters and a preset safety status prediction model; wherein the safety status prediction model adopts a neural network whose main body is a set of weight value matrices. When the neural network learns, the weight of the current layer is updated according to the error function, and the output value of each layer is used as the input value. The updated weight is used for reverse calculation, and the calculated value is used again to calculate the error of the previous layer and update the weight value, and the correlation between the relevant parameters and the battery safety status is introduced to correct the updated weight value.

[0009] Furthermore, the gas content includes the H2, CO, CO2, CH4, C2H6, C2H4, C3H6 and O2 gas contents in the module warehouse.

[0010] Furthermore, the calculation formula for the battery safety status is:

[0011]

[0012] Among them, SOS is the battery safety state; SOS k (x k ) is the safety status value of the relevant parameters, x k is the current value of the relevant parameters; x0 is the completely safe state SOS k =1 when the value of the relevant parameter; x limit The limit value of the acceptable safe state for the relevant parameters.

[0013] Furthermore, when 0≤SOS<tanh(1), it is in a safe state; tanh(1)≤SOS<1-(1-tanh(1)) n When , it is in the warning state; 1-(1-tanh(1)) n When ≤SOS≤1, it is in a dangerous operating state and needs to stop running immediately; where n is the number of related parameters.

[0014] Furthermore, the weight calculation formula of relevant parameters is:

[0015]

[0016] Among them, W is the weight; α is the learning rate; e is the error; SOS kis the safety status value of the input related parameters; SOS is the battery safety status output value obtained by forward calculation of the current input data.

[0017] Furthermore, the error function is:

[0018]

[0019] Where, e is the error; d i is the expected output vector; y oi is the output vector; q, i and k are constants.

[0020] In a second aspect, the present invention further provides a safety status diagnosis system for a lithium iron phosphate battery energy storage system, comprising:

[0021] The data acquisition module is configured to: obtain relevant parameters of the battery, including voltage, current, temperature, deformation, noise, gas content, and dQ / dV-U peak area within a preset voltage range;

[0022] The diagnosis prediction module is configured to obtain the safety status of the battery based on the acquired relevant parameters and a preset safety status prediction model; wherein the safety status prediction model adopts a neural network whose main body is a set of weight value matrices. When the neural network learns, the weight of the current layer is updated according to the error function, and the output value of each layer is used as the input value. The updated weight is used for reverse calculation, and the calculated value is used again to calculate the error of the previous layer and update the weight value, and the correlation between the relevant parameters and the battery safety status is introduced to correct the updated weight value.

[0023] Furthermore, the gas content includes the H2, CO, CO2, CH4, C2H6, C2H4, C3H6 and O2 gas contents in the module warehouse.

[0024] Furthermore, the calculation formula for the battery safety status is:

[0025]

[0026] Among them, SOS is the battery safety state; SOS k (x k ) is the safety status value of the relevant parameters, x k is the current value of the relevant parameters; x0 is the completely safe state SOS k =1 when the value of the relevant parameter; x limit The limit value of the acceptable safe state for the relevant parameters.

[0027] Furthermore, when 0≤SOS<tanh(1), it is in a safe state; tanh(1)≤SOS<1-(1-tanh(1)) nWhen , it is in the warning state; 1-(1-tanh(1)) n When ≤SOS≤1, it is in a dangerous operating state and needs to stop running immediately; where n is the number of related parameters.

[0028] Furthermore, the weight calculation formula of relevant parameters is:

[0029]

[0030] Among them, W is the weight; α is the learning rate; e is the error; SOS k is the safety status value of the input related parameters; SOS is the battery safety status output value obtained by forward calculation of the current input data.

[0031] Furthermore, the error function is:

[0032]

[0033] Where, e is the error; d i is the expected output vector; y oi is the output vector; q, i and k are constants.

[0034] In a third aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the safety status diagnosis method for the lithium iron phosphate battery energy storage system described in the first aspect.

[0035] In a fourth aspect, the present invention further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein when the processor executes the program, the steps of the safety status diagnosis method of the lithium iron phosphate battery energy storage system described in the first aspect are implemented.

[0036] In a fifth aspect, the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps of the safety status diagnosis method of the lithium iron phosphate battery energy storage system described in the first aspect are implemented.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] In the present invention, the safety state of the battery is obtained according to the obtained relevant parameters and the preset safety state prediction model; wherein, the safety state prediction model adopts a neural network whose main body is a group of weight value matrices. When the neural network is learning, the weight of the current layer is updated according to the error function, and the output value of each layer is used as the input value. The updated weight is used for reverse calculation, and the calculated value is used again to calculate the error of the previous layer and update the weight value. The correlation between the relevant parameters and the safety state of the battery is introduced to correct the updated weight value. On the basis of comprehensively considering the voltage, current, temperature, deformation, noise, gas content and dQ / dV-U peak area parameters within the preset voltage range, considering the different degrees of influence of different parameter changes on the thermal runaway of the battery cell, the correlation between the relevant parameters and the safety state of the battery is introduced to correct the updated weight value, thereby improving the correlation between the parameters and the safety state of the battery and ensuring the accuracy of the prediction results.

[0039] Compared with the current battery management methods and systems, the present invention takes into account more comprehensive input parameters, the parameters are more closely related to the electrical safety status, the definition of the battery safety status is more accurate, and the prediction accuracy is higher. The calculation time of the battery cell safety status of the present invention is <0.1s, and the energy storage system can be monitored in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The drawings in the specification that constitute a part of this embodiment are used to provide a further understanding of this embodiment. The schematic embodiments of this embodiment and their descriptions are used to explain this embodiment and do not constitute improper limitations on this embodiment.

[0041] Figure 1 is a flow chart of the method of embodiment 1 of the present invention;

[0042] Figure 2 The neural network structure of embodiment 1 of the present invention;

[0043] Figure 3 is a graph of Example 1 of the present invention with U2 as the abscissa and dQ / dV as the ordinate;

[0044] Figure 4 The temperature, voltage and SOS variation curves of the battery cell during the thermal runaway process of Example 2 of the present invention;

[0045] Figure 5 The actual value and predicted value curve of SOS in the thermal runaway process of Example 2 of the present invention. DETAILED DESCRIPTION

[0046] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0047] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0048] Embodiment 1:

[0049] This embodiment provides a safety status diagnosis method for a lithium iron phosphate battery energy storage system. Figure 1 As shown, one or more battery cells are selected from the module of the energy storage system, and the voltage, current, temperature, deformation, external noise of the battery cells, and the H2, CO, CO2, CH4, C2H6, C2H4, C3H6, and O2 gas contents in the module compartment are monitored through BMU and sensors during the operation of the battery cells. The data collected from the BMU and sensors are transmitted to the cloud in real time through the BMS, where the data is preprocessed and the processed data is passed to the trained deep learning neural network for calculation, which can quickly output the safety status assessment value of the energy storage system, identify the unsafe state of the system early, issue a warning, and respond quickly. The main steps of the method in this embodiment are:

[0050] S1. In this embodiment, optionally, a battery management system is used to obtain voltage, current, center temperature of the large surface of the battery cell, etc., a deformation sensor is used to obtain the large surface deformation of the battery cell, a sound sensor is used to obtain the external noise of the battery cell, and a gas sensor is used to obtain the H2, CO, CO2, CH4, C2H6, C2H4, C3H6 and O2 gas content in the module compartment. All data are obtained at one data point per 1s; among which, the unit obtained by the gas sensor can be set to ppm.

[0051] S2, such as Figure 2 As shown, the safety status prediction model in this embodiment adopts a deep learning neural network, which is a feedforward neural network system based on a multi-layer perception mechanism. The neural network has multiple input parameters, multiple hidden layers and multiple neurons, and can learn in real time according to new data.

[0052] Optionally, the input layer parameters of the deep learning neural network include voltage, current, temperature, deformation, noise, H2, CO, CO2, CH4, C2H6, C2H4, C3H6 and O2 gas content in the module compartment, and dQ / dV-U peak area between the voltage range of 4 and 5V.

[0053] The test method for the dQ / dV-U peak area between the voltage range of 4 and 5 V is as follows: read the voltage, current and time parameters of the battery cell from the BMU, and take a data point every 60 seconds. The calculation method of dQ / dV is as follows:

[0054]

[0055] Where I is the current; t is the sampling interval, which can be 60s; U1 is the voltage at the last sampling point; and U2 is the voltage at the current sampling point. Figure 3 As shown in , the voltage U2 of the current sampling point is used as the horizontal coordinate and dQ / dV is used as the vertical coordinate.

[0056] The peak area of ​​the first characteristic peak between 4V and 5V is calculated by an integration algorithm, and the peak area is one of the input characteristic vectors of the neural network.

[0057] In this embodiment, the deep learning neural network input parameters are converted into values ​​between 0 and 1 through the following normalization formula.

[0058]

[0059] Among them, x' is the normalized data; x is the original data; x min is the minimum value in the data set; x max is the maximum value in the data set.

[0060] In this embodiment, the output parameter of the deep learning neural network is the battery safety state (SOS), which is a value ranging from 0 to 1.

[0061] In this embodiment, optionally, the calculation formula of the battery safety state (SOS) is as follows:

[0062]

[0063] Among them, SOS k (x k ) is the input physical quantity x k The safety status value is calculated as follows:

[0064]

[0065] Among them, x k is the current value of the physical quantity; x0 is SOS k =1 when the value of the physical quantity; x limit is the physical quantity x k The limit of acceptable safety status; SOC k = 1 is a completely safe state. From this we can draw the following conclusions:

[0066] When 0≤SOS k <tanh(1), for the physical quantity x k For example, the system is in a safe state; when tanh(1)≤SOS k ≤1, for the physical quantity x kThe system is in an unsafe state.

[0067] By analogy, when the number of input physical quantities of the early warning system is n, the method for judging the safety status of the system is as follows:

[0068] When 0≤SOS<tanh(1), the system is in a safe state. At this time, the system can be operated safely without intervention.

[0069] tanh(1)≤SOS<1-(1-tanh(1)) n When the system is in a warning state, it means that the system is in a risky operation state. This state means that there are potential safety hazards in the system operation. There will be no thermal safety accidents in continuous operation, but the battery system will be irreversibly damaged. At this time, the background program sends out a warning signal to remind manual or BMS to respond in advance.

[0070] 1-(1-tanh(1)) n When ≤SOS≤1, the system is in a dangerous operating state. Continuous operation will inevitably cause a thermal safety accident. At this time, the background program sends a warning signal and immediately stops the system operation.

[0071] S3. Training of deep learning neural networks:

[0072] The data set required for the deep learning neural network training described in this embodiment includes an input parameter set and an output parameter set. The input parameter set collects a batch of voltage, current, temperature, deformation, noise, H2, CO, CO2, CH4, C2H6, C2H4, C3H6, O2 gas content in the module compartment, and dQ / dV-U peak area parameters between 4 and 5V during the operation of lithium iron phosphate energy storage cells. The output parameter set is the safety status assessment value of the system corresponding to the input parameter, which is obtained by observing and evaluating the system status. The data set contains data under different operating states of the lithium iron phosphate energy storage system, which can truly reflect the operating characteristics of the energy storage system.

[0073] The calculation principle of the deep learning neural network described in this embodiment is as follows:

[0074] S3.1. Initialize the neural network:

[0075] During the first training, a random number is used to initialize the neural network. If a trained neural network has already been used, it can be used to initialize the neural network. The main body of the neural network is a set of weight value matrices W. Here, the number of layers of the neural network and the number of neural nodes in each layer need to be defined. For example, a neural network with n1, n2, n3 and n4 neural nodes is defined. The network has four layers in total, and three weight matrices are initialized, with dimensions n1+1×n2+1, n2+1×n3+1 and n3+1×n4. Except for the last output layer, a bias value b is added to each layer to make the neural network converge quickly.

[0076] S3.2, forward propagation:

[0077] The input value δ of the i-th neuron in the j-th layer is forward propagated j , i Equal to the value of all neurons in the j-1th layer δ j-1 , i Multiply it by the connection weight w of the i-th neuron in the j-th layer j-1 , i The sum of the products of and the bias of the i-th neuron in the j-th layer is formed; assuming that there are K neurons in the j-1-th layer and M neurons in the j-th layer, that is:

[0078]

[0079] The output value of each layer is:

[0080] a j，i (x) = active(δ j，i (x)

[0081] Among them, active() is the activation function, which is used to perform nonlinear classification on data. The options are sigmod, tanh and relu.

[0082] The input vector of the output layer is y i =(y i1 ,y i2 , …, y iq ), the output vector is y o =(y o1 ,y o2 ,…y oq ), the expected output vector is d o =(d1, d2, ..., d q ).

[0083] The calculated error function is:

[0084]

[0085] S3.3, Error back propagation:

[0086] Update the weights of the current layer according to the error function:

[0087]

[0088] Among them, α is the learning rate.

[0089] Take the output value of each layer as the input value and use the updated weights to perform reverse calculations:

[0090]

[0091] a j-1，i (x) = active(δ j-1，i (x)

[0092] The calculated value is used again to calculate the error e of the previous layer and update the weight value, thereby completing the update of all weight values, which is called one-step learning.

[0093] During the entire training process, you can define the number of learning times or limit the threshold of the error function to make the model converge.

[0094] S3.4. Use the trained neural network for prediction:

[0095] Input the input vector of the predicted data, perform a forward propagation calculation, and the output vector y of the output layer o =(y o1 ,y o2 ,…y oq ) is the predicted value.

[0096] S3.5. The influence of changes in different physical quantities on thermal runaway of the battery cell is different. Therefore, in this embodiment, when training the neural network, the correlation between the physical quantity and the battery safety state (SOS) is introduced to correct the learning process of the neural network. The specific operation is as follows:

[0097] The correlation between the physical quantity and the battery safety state (SOS) can be expressed by the correlation coefficient, which is defined as follows:

[0098]

[0099] Among them, SOS k The input physical quantity x k The safety status value of SOS; SOS is the battery safety status output value obtained by forward calculation of the current input data; when the neural network performs reverse conduction calculation, the weight calculation formula of this physical quantity is modified as follows:

[0100]

[0101] Therefore, when training the neural network, the learning weights of physical quantities that are highly correlated with the system battery state of safety (SOS) are strengthened, while the learning weights of physical quantities that are less correlated with the system are weakened, which is conducive to learning a neural network with higher accuracy.

[0102] Embodiment 2:

[0103] In this embodiment, the battery cells in the energy storage system module are operated under the following conditions: the SOC of the battery cells in the initial state is 100%, the charging current of the battery cells is 0.5C, and a 1000W heating plate is attached to the large surface of the battery cells. During the test, the safety state diagnosis method of the lithium iron phosphate battery energy storage system in Example 1 is used to predict the safety state of the battery cells. The voltage, temperature, SOS, etc. of the battery cells are Figure 4 shown.

[0104] Figure 4 In the test, as the test progressed, the system predicted that the SOS of the battery cell gradually increased from 0.38 to 0.99, reflecting the whole process of the battery cell losing control. Figure 4 As shown in the midpoint ①, when the time is less than 100s, SOS<0.762(tanh(1)). During this period, the battery cell is in a safe operating state and no irreversible reaction occurs inside the battery cell. Figure 4 As shown in the middle point ②, when the time is greater than 100s and less than 440s, the cell SOS is between 0.762 and 0.997 (1-(1-tanh(1)) 4 ), indicating that the battery is in a risky operating state. Although the battery has not experienced thermal runaway, an irreversible reaction has occurred inside it and the battery has failed. When the time is > 440s, the battery SOS> 0.997, and the battery is in a dangerous operating state. Continuous operation will cause thermal runaway. The design test value shows that if Figure 4 As shown at the midpoint, the valve opening time of the battery cell is 540s, so the algorithm of the present invention issues a warning 440s earlier than the battery cell valve opening.

[0105] In this example, the SOS predicted value by the proposed neural network algorithm and the actual value calculated by the formula are compared. Figure 5 As shown, the square root error of the predicted value is 2.83% compared to the actual value.

[0106] Embodiment 3:

[0107] This embodiment provides a safety status diagnosis system for a lithium iron phosphate battery energy storage system, including:

[0108] The data acquisition module is configured to: obtain relevant parameters of the battery, including voltage, current, temperature, deformation, noise, gas content, and dQ / dV-U peak area within a preset voltage range;

[0109] The diagnosis prediction module is configured to obtain the safety status of the battery based on the acquired relevant parameters and a preset safety status prediction model; wherein the safety status prediction model adopts a neural network whose main body is a set of weight value matrices. When the neural network learns, the weight of the current layer is updated according to the error function, and the output value of each layer is used as the input value. The updated weight is used for reverse calculation, and the calculated value is used again to calculate the error of the previous layer and update the weight value, and the correlation between the relevant parameters and the battery safety status is introduced to correct the updated weight value.

[0110] Optionally, the gas content includes H2, CO, CO2, CH4, C2H6, C2H4, C3H6 and O2 gas contents in the module warehouse.

[0111] Optionally, the battery safety status is calculated as:

[0112]

[0113] Among them, SOS is the battery safety state; SOS k (x k ) is the safety status value of the relevant parameters, x k is the current value of the relevant parameters; x0 is the completely safe state SOS k =1 when the value of the relevant parameter; x limit It is the limit value of the safe state that the relevant parameters can accept. When 0≤SOS<tanh(1), it is in a safe state; tanh(1)≤SOS<1-(1-tanh(1)) n When , it is in the warning state; 1-(1-tanh(1)) n When ≤SOS≤1, it is in a dangerous operating state and needs to stop running immediately; where n is the number of related parameters.

[0114] Optionally, the weight calculation formula for related parameters is:

[0115]

[0116] Among them, W is the weight; α is the learning rate; e is the error; SOS k is the safety status value of the input related parameters; SOS is the battery safety status output value obtained by forward calculation of the current input data.

[0117] Optionally, the error function is:

[0118]

[0119] Where, e is the error; d i is the expected output vector; y oi is the output vector; q, i and k are constants.

[0120] The working method of the system is the same as the safety status diagnosis method of the lithium iron phosphate battery energy storage system in Example 1, and will not be repeated here.

[0121] Embodiment 4:

[0122] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the safety status diagnosis method of the lithium iron phosphate battery energy storage system described in Example 1 are implemented.

[0123] Embodiment 5:

[0124] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor. When the processor executes the program, the steps of the safety status diagnosis method of the lithium iron phosphate battery energy storage system described in Example 1 are implemented.

[0125] Embodiment 6:

[0126] This embodiment provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps of the safety status diagnosis method of the lithium iron phosphate battery energy storage system described in Example 1 are implemented.

[0127] The above description is only a preferred embodiment of the present embodiment and is not intended to limit the present embodiment. For those skilled in the art, the present embodiment may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present embodiment shall be included in the protection scope of the present embodiment.

Claims

1. A safety status diagnosis method for a lithium iron phosphate battery energy storage system, characterized in that: include: Obtain relevant parameters of the battery, including voltage, current, temperature, deformation, noise, gas content, and dQ / dV-U peak area within a preset voltage range; The safety status of the battery is obtained based on the obtained relevant parameters and a preset safety status prediction model; wherein the safety status prediction model adopts a neural network whose main body is a set of weight value matrices. When the neural network learns, the weight of the current layer is updated according to the error function, and the output value of each layer is used as the input value. The updated weight is used for reverse calculation, and the calculated value is used again to calculate the error of the previous layer and update the weight value, and the correlation between the relevant parameters and the battery safety status is introduced to correct the updated weight value.

2. The safety status diagnosis method of the lithium iron phosphate battery energy storage system according to claim 1, characterized in that: The gas content includes the H2, CO, CO2, CH4, C2H6, C2H4, C3H6 and O2 gas contents in the module warehouse.

3. The safety status diagnosis method of the lithium iron phosphate battery energy storage system according to claim 1, characterized in that: The calculation formula for the battery safety status is: Among them, SOS is the battery safety state; SOS k (x k ) is the safety status value of the relevant parameters, x k is the current value of the relevant parameters; x0 is the completely safe state SOS k =1 when the value of the relevant parameter; x limit The limit value of the acceptable safe state for the relevant parameters.

4. The safety status diagnosis method of the lithium iron phosphate battery energy storage system according to claim 3, characterized in that: When 0≤SOS<tanh(1), it is in a safe state; tanh(1)≤SOS<1-(1-tanh(1)) n When , it is in the warning state; 1-(1-tanh(1)) n When ≤SOS≤1, it is in a dangerous operating state and needs to stop running immediately; where n is the number of related parameters.

5. The safety status diagnosis method of the lithium iron phosphate battery energy storage system according to claim 1, characterized in that: The weight calculation formula of relevant parameters is: Among them, W is the weight; α is the learning rate; e is the error; SOS k is the safety status value of the input related parameters; SOS is the battery safety status output value obtained by forward calculation of the current input data.

6. The safety status diagnosis method of the lithium iron phosphate battery energy storage system according to claim 1, characterized in that: The error function is: Where, e is the error; d i is the expected output vector; y oi is the output vector; q, i and k are constants.

7. A safety status diagnostic system for a lithium iron phosphate battery energy storage system, characterized in that: include: The data acquisition module is configured to: obtain relevant parameters of the battery, including voltage, current, temperature, deformation, noise, gas content, and dQ / dV-U peak area within a preset voltage range; The diagnosis prediction module is configured to obtain the safety status of the battery based on the acquired relevant parameters and a preset safety status prediction model; wherein the safety status prediction model adopts a neural network whose main body is a set of weight value matrices. When the neural network learns, the weight of the current layer is updated according to the error function, and the output value of each layer is used as the input value. The updated weight is used for reverse calculation, and the calculated value is used again to calculate the error of the previous layer and update the weight value, and the correlation between the relevant parameters and the battery safety status is introduced to correct the updated weight value.

8. The safety status diagnostic system of the lithium iron phosphate battery energy storage system according to claim 7, characterized in that: The gas content includes the H2, CO, CO2, CH4, C2H6, C2H4, C3H6 and O2 gas contents in the module warehouse.

9. The safety status diagnostic system of the lithium iron phosphate battery energy storage system according to claim 7, characterized in that: The calculation formula for the battery safety status is: Among them, SOS is the battery safety state; SOS k (x k ) is the safety status value of the relevant parameters, x k is the current value of the relevant parameters; x0 is the completely safe state SOS k =1 when the value of the relevant parameter; x limit The limit value of the acceptable safe state for the relevant parameters.

10. The safety status diagnostic system of the lithium iron phosphate battery energy storage system according to claim 9, characterized in that: When 0≤SOS<tanh(1), it is in a safe state; tanh(1)≤SOS<1-(1-tanh(1)) n When , it is in the warning state; 1-(1-tanh(1)) n When ≤SOS≤1, it is in a dangerous operating state and needs to stop running immediately; where n is the number of related parameters.

11. The safety status diagnosis method of the lithium iron phosphate battery energy storage system according to claim 1, characterized in that: The weight calculation formula of relevant parameters is: Among them, W is the weight; α is the learning rate; e is the error; SOS k is the safety status value of the input related parameters; SOS is the battery safety status output value obtained by forward calculation of the current input data.

12. The safety status diagnostic system of the lithium iron phosphate battery energy storage system according to claim 7, characterized in that: The error function is: Where, e is the error; d i is the expected output vector; y oi is the output vector; q, i and k are constants.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the safety status diagnosis method of the lithium iron phosphate battery energy storage system as described in any one of claims 1 to 6 are implemented.

14. An electronic device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that: When the processor executes the program, the steps of the safety status diagnosis method of the lithium iron phosphate battery energy storage system as described in any one of claims 1-6 are implemented.

15. A computer program product, characterized in that The computer program product includes a computer program, and when the computer program is executed by a processor, the steps of the safety status diagnosis method of the lithium iron phosphate battery energy storage system as described in any one of claims 1 to 6 are implemented.

Images