Real-time SOC prediction method for all-vanadium redox flow battery system

Abstract

The present application discloses a real-time SOC prediction method for an all-vanadium redox flow battery system, comprising: sampling positive and negative electrolytes in a charge and discharge cycle of an all-vanadium redox flow battery system to determine the concentrations of positive electrode tetravalent and pentavalent vanadium ions and the concentrations of negative electrode divalent and trivalent vanadium ions; recording charge and discharge parameters at the moment of sampling the positive and negative electrolytes in the charge and discharge cycle; taking the charge and discharge parameters as feature vectors X; taking the concentrations of the positive electrode tetravalent and pentavalent vanadium ions and the concentrations of the negative electrode divalent and trivalent vanadium ions as target functions y to establish a database; dividing data in the database into a training set and a test set; and using an artificial intelligence algorithm to train the data in the training set, and building a prediction model for the concentrations of the positive electrode tetravalent and pentavalent vanadium ions and the concentrations of the negative electrode divalent and trivalent vanadium ions of the all-vanadium redox flow battery. The present method can predict in real time the concentrations of active materials in the positive and negative electrolytes of the all-vanadium redox flow battery system in the charge and discharge cycle, and realize SOC real-time monitoring of the all-vanadium redox flow battery system.

Description

A real-time SOC prediction method for all-vanadium liquid flow battery system Technical Field

[0001] The present application relates to a real-time SOC prediction method for an all-vanadium liquid flow battery system, belonging to the technical field of all-vanadium liquid flow batteries. Background Art

[0002] Monitoring the state of charge (SOC) of vanadium flow battery systems is crucial for system safety and stability. A reliable real-time SOC prediction system for vanadium flow battery systems can significantly improve system stability and reduce operation and maintenance costs. Currently, methods for estimating SOC for vanadium flow battery systems primarily include discharge experiments, ampere-hour integration, open-circuit voltage, and Kalman filtering. The difficulty in predicting the real-time SOC of vanadium flow battery systems lies in the fact that electrolyte migration during operation can cause an imbalance between the positive and negative electrolytes, rendering the real-time SOC estimate inaccurate. The discharge experiment method offers high accuracy but cannot be used for online measurement; the ampere-hour integration and open-circuit voltage methods cannot account for the accuracy loss caused by electrolyte migration; and the Kalman filter method relies heavily on the accuracy of the battery model and cannot account for the impact of electrolyte migration during vanadium flow battery operation. Therefore, none of these methods can achieve real-time SOC prediction for large-scale demonstration and commercial vanadium flow battery systems.

[0003] Summary of the Invention

[0004] According to one aspect of the present application, a real-time SOC prediction method for an all-vanadium liquid flow battery system is provided. In view of the limitations and complexity of existing SOC detection methods, the present application only utilizes the charge and discharge parameter data of the all-vanadium liquid flow battery system during the charge and discharge process, and realizes real-time online prediction of SOC without introducing additional complex acquisition instruments and measurement equipment, thereby solving the problem that existing SOC detection methods are difficult to achieve both high precision and real-time performance.

[0005] This application adopts the following technical solutions:

[0006] A real-time SOC prediction method for an all-vanadium liquid flow battery system comprises the following steps:

[0007] S1. Perform a charge-discharge cycle test on the all-vanadium redox flow battery system, sample the electrolyte in the electrolyte storage tank several times during the charge-discharge cycle, obtain the vanadium ion concentration in the electrolyte, and record the charge-discharge parameters at the sampling time;

[0008] S2. Using the charge and discharge parameters recorded in step S1 as the characteristic vector X and the vanadium ion concentration in the electrolyte obtained in step S1 as the objective function y, establishing a database, dividing the data in the database into a training set and a test set, and using an artificial intelligence algorithm to train and model the data in the training set. After data verification and model parameter adjustment, a prediction model for predicting the vanadium ion concentration in the electrolyte is obtained;

[0009] S3. When the all-vanadium liquid flow battery system is running, the charge and discharge parameters are collected in real time, and the real-time vanadium ion concentration in the electrolyte is predicted by the prediction model in step S2. Then, the real-time SOC of the all-vanadium liquid flow battery system is calculated according to formula (1):

[0010] In the formula, [V 2+ ]、[V 3+ ]、[VO 2+ ]and represent the concentrations of divalent, trivalent, tetravalent and pentavalent vanadium ions, respectively.

[0011] Optionally, in step S1, obtaining the vanadium ion concentration in the electrolyte includes:

[0012] The concentration of tetravalent and pentavalent vanadium ions in the positive electrode electrolyte storage tank is obtained, and / or the concentration of divalent and trivalent vanadium ions in the negative electrode electrolyte storage tank is obtained.

[0013] Optionally, in step S1, the charge and discharge parameter is selected from at least one of current, voltage, average shelf voltage, charge capacity during charging, and discharge capacity during discharging.

[0014] Optionally, in step S1, the charge-discharge cycle includes n groups of charge-discharge cycles, n>1;

[0015] The sampling process includes:

[0016] During the charging cycle, at least 4 samples were taken, including the first sample within 1 minute after the start of charging and the last sample within 1 minute before the end of charging.

[0017] During the discharge cycle, at least four samples were taken, including the first sample within 1 minute after the start of charging and the last sample within 1 minute before the end of charging.

[0018] Optionally, in step S1, the process of obtaining the vanadium ion concentration in the electrolyte after sampling is as follows: detecting the vanadium ion concentration in the electrolyte by titration. In step S1, the process of obtaining the vanadium ion concentration in the electrolyte after sampling is as follows: detecting the vanadium ion concentration in the electrolyte by titration, and for the vanadium ion concentration corresponding to the moment when no sampling is performed, the vanadium ion concentration of two adjacent sampling titrations and the capacity difference between charge and discharge are calculated by interpolation method according to formula (2):

[0019] Where c(t1), c(t2), c(t i ) represent the vanadium ion concentrations at the first sampling point, the second sampling point and time ti, respectively. C(t1), C(t2), C(t i ) represent the charge or discharge capacity at the first sampling point, the second sampling point and the moment ti, respectively.

[0020] Optionally, in step S2, a database is established using the n consecutive groups of charge and discharge parameters recorded in step S1 as a feature vector X and the vanadium ion concentration in the electrolyte corresponding to the nth group of charge and discharge cycles as an objective function y.

[0021] Optionally, the data of any m1 group of cycles in the database is used as a training set of charging cycles;

[0022] The data of any z1 group of cycles in the database is used as the test set of charging cycles;

[0023] The data of any m2 group of cycles in the database is used as the training set of discharge cycles;

[0024] The data of any z2 groups of cycles in the database are used as the test set of discharge cycles;

[0025] Wherein, m1, z1, m2, and z2 are independently smaller than n.

[0026] Optionally, in step S2, the artificial intelligence algorithm is selected from one of the autoregressive integrated moving average algorithm (ARIMA), Prophet algorithm, gradient boosting machine algorithm (GBM), Seq2Seq algorithm, TCN algorithm, WaveNet algorithm, Informer algorithm, and Transformer algorithm.

[0027] Optionally, the Seq2Seq algorithm is selected from one of a recurrent neural network RNN, a long short-term memory neural network LSTM, and an autoregressive recurrent neural network DeepAR.

[0028] Optionally, when the artificial intelligence algorithm is a long short-term memory neural network (LSTM) algorithm, the activation function is "relu", the neural network optimizer is "Adam", the number of hidden layers is at least 1, the number of neurons in each layer is at least 8, and a part of the neurons are randomly discarded during the training process to prevent overfitting.

[0029] Optionally, in step S2, the data modeling process further includes adopting a regularization method to prevent overfitting;

[0030] The regularization method is selected from one of L1 regularization, L2, Dropout, or a composite regularization method of at least any two.

[0031] Optionally, the parameter of the regularization method is (0, 1].

[0032] Optionally, neurons are randomly dropped from the range (0,1).

[0033] Optionally, the prediction model for predicting the concentration of vanadium ions in the electrolyte is a concentration prediction model for tetravalent and pentavalent vanadium ions, or a concentration prediction model for divalent and trivalent vanadium ions.

[0034] The beneficial effects of this application include:

[0035] A real-time SOC prediction method for all-vanadium redox flow battery systems uses parameters such as current, voltage, and charge / discharge capacity during the charge and discharge processes of all-vanadium redox flow batteries as input features. Using artificial intelligence algorithms, a prediction model for the concentration of vanadium ions in each valence state at the positive and negative electrodes is established. This method utilizes real-time data from the charge and discharge processes to perform online predictions of the electrolyte SOC of all-vanadium redox flow batteries after positive and negative electrode imbalance. The predictions are sensitive and accurate. This method is low-cost, requires no additional data acquisition devices or testing instruments, and is easy to implement and maintain. The SOC prediction model for the positive electrode of all-vanadium redox flow batteries, developed using parameter data from the charge and discharge processes of individual cells, can be applied to large-scale demonstration and commercial systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] FIG1 is a flow chart showing the implementation steps of the real-time SOC prediction method for the all-vanadium liquid flow battery system of the present application.

[0037] Figure 2 is a comparison chart of the model prediction results and actual values ​​of the tetravalent and pentavalent vanadium ion concentrations at the positive electrode during the charging process in Example 1 of the present application.

[0038] FIG3 is a comparison chart of the model prediction results and the actual values ​​of the tetravalent and pentavalent vanadium ion concentrations at the positive electrode during the discharge process in Example 1 of the present application. DETAILED DESCRIPTION

[0039] The present application is described in detail below with reference to embodiments, but the present application is not limited to these embodiments.

[0040] Unless otherwise specified, the raw materials in the examples of this application were purchased through commercial channels.

[0041] Unless otherwise specified, conventional methods were used for testing, and instrument settings were those recommended by the manufacturer.

[0042] Example 1

[0043] As shown in the flow of the implementation steps in FIG1 , a method for predicting the real-time SOC of the all-vanadium flow battery system is implemented using a prediction model of the tetravalent and pentavalent vanadium ion concentrations at the positive electrode of the all-vanadium flow battery system as follows:

[0044] Step 1: Perform a charge-discharge test on the all-vanadium redox flow battery system. Samples of the positive and negative electrolytes during the test are titrated to determine the concentrations of tetravalent and pentavalent vanadium ions at the positive electrode, and the concentrations of divalent and trivalent vanadium ions at the negative electrode. In this example, the SOC of the all-vanadium redox flow battery system is determined based on the concentrations of tetravalent and pentavalent vanadium ions at the positive electrode. The system's rated power is 5 kW. The charge and discharge test process is as follows: first, place it for 30 seconds, then perform 5kW constant power charging, charge to a cut-off voltage of 1.55V, then switch to constant voltage charging of 1.53V, charge to 100% SOC, and place it for 30 seconds; finally, perform constant power discharge, discharge to a cut-off voltage of 1V, and complete one charge and discharge cycle; the titration sampling time is as follows: for one charging cycle, take at least 4 samples, take 1 sample within 1 minute after the start of charging, take 1 sample within 1 minute before the end of charging, and take at least 2 more samples during the charging process; for one discharge cycle, take at least 4 samples, take 1 sample within 1 minute after the start of discharge, take a sample within the last minute before the end of discharge, and take at least 2 more samples during the discharge process.

[0045] Step 2: Based on the titrated tetravalent and pentavalent vanadium ion concentrations of the positive electrode, according to the charge and discharge capacity, the tetravalent and pentavalent vanadium ion concentrations of the positive electrode in the cycle are interpolated and calculated; for the vanadium ion concentration corresponding to the moment when no sampling is performed, according to the vanadium ion concentrations of the two adjacent sampling titrations and the charge and discharge capacity difference, the vanadium ion concentration is calculated by interpolation method according to formula (2):

[0046] Where c(t1), c(t2), c(t i ) represent the vanadium ion concentrations at the first sampling point, the second sampling point and time ti, respectively. C(t1), C(t2), C(t i ) represent the charge or discharge capacity at the first sampling point, the second sampling point and the moment ti, respectively.

[0047] A database is established using the characteristics of n consecutive charge-discharge cycles as input feature X, and the tetravalent and pentavalent vanadium ion concentrations of the positive electrode corresponding to the nth charge-discharge cycle as the target function y. The charge-discharge cycle characteristics include current, voltage, average shelf voltage, charge capacity (during charging), and discharge capacity (during discharging); n is an integer greater than or equal to 1.

[0048] Step 3: Divide the data in the database into training sets and test sets. In this example, for charging cycles, the data of the 9th, 15th, 55th, and 81st charging cycles are used as training sets, and the data of the 88th charging cycle is used as the test set. For discharging cycles, the data of the 2nd, 21st, 29th, and 39th discharge cycles are used as training sets, and the data of the 61st discharge cycle is used as the test set.

[0049] Step 4: An artificial intelligence algorithm is used to train the data in the training set to establish a prediction model for the concentration of tetravalent and pentavalent vanadium ions in the positive electrode during the charging and discharging process of the all-vanadium liquid flow battery. The artificial intelligence uses a long short-term memory neural network (LSTM) algorithm with 1 hidden layer, 32 neurons, "relu" activation function, and "Adam" neural network optimizer. The L2 regularization method is used to prevent overfitting with a parameter of 0.01. 20% of neurons are randomly discarded during the training process to prevent overfitting. The predicted and actual results of the tetravalent and pentavalent vanadium ion concentrations in the positive electrode are shown in Figure 2, respectively, and the model training accuracy is shown in Table 1.

[0050] Step 5: Use the data in the test set to evaluate the trained prediction model for tetravalent and pentavalent vanadium ions in the positive electrode of the all-vanadium flow battery. The training results are shown in Figure 2, and the accuracy of the test set is shown in Table 1.

[0051] Step 6: When the all-vanadium liquid flow battery system is running, the charge and discharge parameter data are collected in real time, and the real-time vanadium ion concentration in the positive electrode electrolyte is predicted by the positive electrode tetravalent and pentavalent vanadium ion prediction model, and then the real-time SOC of the all-vanadium liquid flow battery system is calculated according to formula (1):

[0052] In the formula, [V 2+ ]、[V 3+ ]、[VO 2+ ]and represent the concentrations of divalent, trivalent, tetravalent and pentavalent vanadium ions, respectively.

[0053] Table 1 Model accuracy of the prediction model for tetravalent vanadium and pentavalent vanadium in the cathode of all-vanadium liquid flow battery system

[0054] The above descriptions are merely a few embodiments of the present application and do not constitute any form of limitation to the present application. Although the present application discloses the preferred embodiments as above, they are not intended to limit the present application. Any technical personnel familiar with the present profession, without departing from the scope of the technical solution of the present application, using the technical content disclosed above to make slight changes or modifications are equivalent to equivalent implementation cases and fall within the scope of the technical solution.

Claims

1. A real-time SOC prediction method for an all-vanadium liquid flow battery system, characterized in that: The steps include: S1. Perform a charge-discharge cycle test on the all-vanadium liquid flow battery system, sample the electrolyte in the electrolyte storage tank several times during the charge-discharge cycle, obtain the vanadium ion concentration in the electrolyte, and record the charge-discharge parameters at the sampling time; S2, using the charge and discharge parameters recorded in step S1 as the feature vector X, and using the vanadium ion concentration in the electrolyte obtained in step S1 as the objective function y to establish a database, dividing the data in the database into a training set and a test set, using an artificial intelligence algorithm to train and model the data in the training set, and after data verification and model parameter adjustment, obtaining a prediction model for predicting the vanadium ion concentration in the electrolyte; S3, when the all-vanadium liquid flow battery system is running, the charging and discharging parameters are collected in real time, and the real-time vanadium ion concentration in the electrolyte is predicted by the prediction model in step S2, and then the real-time SOC of the all-vanadium liquid flow battery system is calculated according to formula (1): In the formula, [V 2+ ]、[V 3+ ]、[VO 2+ ]and Represent the concentrations of divalent, trivalent, tetravalent and pentavalent vanadium ions, respectively.

2. The real-time SOC prediction method for an all-vanadium liquid flow battery system according to claim 1, characterized in that: In step S1, obtaining the vanadium ion concentration in the electrolyte includes: The concentration of tetravalent and pentavalent vanadium ions in the positive electrode electrolyte storage tank is obtained, and / or the concentration of divalent and trivalent vanadium ions in the negative electrode electrolyte storage tank is obtained.

3. The real-time SOC prediction method for an all-vanadium liquid flow battery system according to claim 1, characterized in that: In step S1, the charge and discharge parameter is selected from at least one of current, voltage, average voltage during storage, charge capacity during charging, and discharge capacity during discharge.

4. The real-time SOC prediction method for an all-vanadium liquid flow battery system according to claim 1, characterized in that: In step S1, the charge-discharge cycle includes n groups of charge-discharge cycles, n>1; The sampling process includes: During the charging cycle, at least 4 samplings are performed, including the first sampling within 1 minute after the start of charging and the last sampling within 1 minute before the end of charging; During the discharge cycle, at least 4 samplings were performed, including the first sampling within 1 minute after the start of charging and the last sampling within 1 minute before the end of charging.

5. The real-time SOC prediction method for an all-vanadium liquid flow battery system according to claim 1, characterized in that: In step S1, the process of obtaining the vanadium ion concentration in the electrolyte after sampling is as follows: the vanadium ion concentration in the electrolyte is detected by titration, and the vanadium ion concentration corresponding to the moment when no sampling is performed is calculated by interpolation method according to the vanadium ion concentration of two adjacent sampling titrations and the capacity difference of charge and discharge according to formula (2): In the formula, c(t1), c(t2), c(t i ) represent the vanadium ion concentrations at the first sampling point, the second sampling point and the time ti, respectively. C(t1), C(t2), C(t i ) represent the charge or discharge capacity at the first sampling point, the second sampling point and the moment ti, respectively.

6. The real-time SOC prediction method for an all-vanadium liquid flow battery system according to claim 4, characterized in that: In step S2, a database is established with the n consecutive groups of charge and discharge parameters recorded in step S1 as the characteristic vector X and the vanadium ion concentration in the electrolyte corresponding to the nth group of charge and discharge cycles as the target function y.

7. The real-time SOC prediction method for an all-vanadium liquid flow battery system according to claim 6, characterized in that: The data of any m1 group of cycles in the database is used as the training set of charging cycles; Take any z1 group of cycle data in the database as the test set of charging cycles; The data of any m2 group cycle in the database is used as the training set of discharge cycles; Take any z2 group of cycle data in the database as the test set of discharge cycles; Among them, m1, z1, m2, z2 are independently less than n.

8. The real-time SOC prediction method for an all-vanadium liquid flow battery system according to claim 1, characterized in that: In step S2, the artificial intelligence algorithm is selected from one of an autoregressive integrated moving average algorithm, a Prophet algorithm, a gradient boosting algorithm, a Seq2Seq algorithm, a TCN algorithm, a WaveNet algorithm, an Informer algorithm, and a Transformer algorithm; The Seq2Seq algorithm is selected from one of a recurrent neural network RNN, a long short-term memory neural network LSTM, and an autoregressive recurrent neural network DeepAR.

9. The real-time SOC prediction method for an all-vanadium liquid flow battery system according to claim 1, characterized in that: In step S2, the data modeling process further includes using a regularization method to prevent overfitting; The regularization method is selected from one of L1 regularization, L2, Dropout, or a composite regularization method of at least any two of them.

10. The real-time SOC prediction method for an all-vanadium liquid flow battery system according to claim 2, characterized in that: The prediction model for predicting the concentration of vanadium ions in the electrolyte is a concentration prediction model for tetravalent and pentavalent vanadium ions, or a concentration prediction model for divalent and trivalent vanadium ions.

Images