A method for estimating the remaining service life of lithium batteries based on load pattern recognition

Abstract

The present invention relates to a method for estimating the remaining service life of a lithium battery based on load mode recognition, which comprises the following steps: load operation superposition mode recognition: sampling data of the lithium battery, and identifying the load as being in one of a constant current mode, a constant resistance mode, a constant power mode, and a superposition mode according to the sampled data; prediction of the remaining service life of the lithium battery in the superposition mode: constructing an SOC table dependent on voltage and current, and calculating the energy capacity Q after the k△t moment according to the current energy capacity value Q0, the voltage V0, and the SOC table. k and voltage V k , when k=M, SOC occurs M ≤0, then the remaining usage time is M△t. This method for estimating the remaining usage time of lithium batteries based on load pattern recognition identifies the load pattern. Based on the load pattern determination, it borrows historical data and uses differential calculus and difference methods to solve nonlinear optimization problems, thereby more accurately estimating the remaining operating time of lithium batteries.

Description

Technical Field

[0001] The invention relates to a method for estimating the remaining service time of a lithium battery based on load mode recognition. Background Art

[0002] Estimating the remaining battery life during use is directly related to battery usage decisions and plans. Accurately estimating the remaining time, based on load conditions, has long been a focus of research and a topic for both industry and academics. Literature searches reveal a relatively large number of papers examining overall battery life prediction, while relatively few focus on the remaining time or range of a battery during a single use. Furthermore, the technical content and accuracy of existing estimation methods are limited, primarily due to the difficulty in calculating the actual capacity of the core technology. Given the practical needs of users, developing a theoretically sound method for accurately predicting the remaining battery life has significant practical value. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a method for estimating the remaining service life of a lithium battery based on load pattern recognition. The method identifies the load pattern, and on the basis of the load pattern determination, borrows historical data, utilizes differential calculus and difference methods to solve nonlinear optimization problems, and thus more accurately estimates the remaining operating time of the lithium battery.

[0004] The method for estimating the remaining service life of a lithium battery based on load pattern recognition of the present invention comprises the following steps:

[0005] S1-Load operation superposition mode recognition: the total voltage V, total current I, and energy capacity Q of the lithium battery are sampled. According to the sampling data, the total voltage V k , total current I k , energy capacity Q k , identifying the load as operating in one of the constant current mode, constant resistance mode, constant power mode, and superposition mode;

[0006] S2-Prediction of the remaining time of lithium battery in superposition mode: Construct an SOC table that depends on voltage and current, and calculate the energy capacity Q after k△t time based on the current energy capacity value Q0, voltage V0, and SOC table k and voltage V k , when k=M, SOC occurs M ≤0, the remaining usage time is M△t.

[0007] Furthermore, the load operation superposition mode identification in step S1 is specifically as follows:

[0008] The time step of data sampling is Δt, and m+1 data records are taken at the current moment. The total voltage V k , total current Ik , energy capacity Q k (k=0,1,…,m), the virtual resistance is The total load power is

[0009] Calculation sequence I k ,R k ,W k Mean difference and standard deviation S I =std(I k ),S R =std(R k ),W I =std(W k );

[0010] like The load is identified as constant current mode, and the current is like It is identified as constant power mode, and the power is like The load is identified as constant resistance mode, and the resistance is Here It is a threshold value specified based on the data level and experimental experience. If the load does not belong to one of the three single modes mentioned above, the load is identified as a superposition mode.

[0011] In the superposition mode, assuming the load is current I, resistance R, and power W, the energy consumed in the kth time period is The actual energy consumption is E k =Q k -Q k-1 , the difference between the two is The total sum of squared errors is Obtaining the optimal solution through nonlinear numerical optimization methods Right now The load pattern can thus be viewed as a current resistance and power A superimposed combination of three parts.

[0012] Furthermore, the remaining time of the lithium battery in the superposition mode in step S2 is specifically predicted as follows:

[0013] Assume the voltage range of lithium battery is [V min ,V max ], the current range is [I min ,I max ]; divide the voltage interval m into equal parts, and the voltage step length is The point is V k =V min +kV h,k=0,1,…,m; divide the current interval n into equal current steps Point I j =I min +jI h ,j=0,1,…,n; to construct the SOC table SOC that depends on voltage and current kj =F(V k ,I j ), k=0,1,…,m; j=0,1,…,n;

[0014] The actual energy capacity Q of the known battery c , energy capacity offset Q d , the current energy capacity value is Q0, the voltage is V0; suppose the load mode is constant current I c , constant resistance R c , constant power W c The energy consumed by the load running for Δt is The energy after Δt is Q1=Q0-Q Δ , SOC value is

[0015] If SOC1>0, calculate the total current after Δt From SOC1 and I1, the voltage value V1 after the first Δt moment can be calculated using the SOC table; and by analogy, the battery capacity and voltage Q after the kth Δt moment can be obtained. k ,V k , the energy capacity and voltage after (k+1)Δt can be calculated, specifically: Further calculation and the total current By SOC k+1 ,I k+1 Use the SOC table to calculate the voltage V after (k+1)Δt k+1 ; Until the calculation reaches step M, SOC occurs M ≤0, the estimated remaining time is MΔt.

[0016] The present invention discloses a method for estimating the remaining service life of a lithium battery based on load pattern recognition. By identifying the load pattern and, based on the load pattern determination, using historical data, differential calculus and difference methods to solve nonlinear optimization problems, the remaining service life of the lithium battery can be estimated more accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The following is a further description of a method for estimating the remaining service life of a lithium battery based on load pattern recognition according to the present invention with reference to the accompanying drawings:

[0018] Figure 1This is a specific implementation flow chart of the lithium battery remaining time prediction in the superposition mode described in Implementation 3 of the lithium battery remaining service time estimation method based on load pattern recognition;

[0019] Figure 2 This is a comparison chart of the used time and the estimated total usage time in an embodiment of the method for estimating the remaining usage time of a lithium battery based on load pattern recognition;

[0020] Figure 3 This is a linear relationship diagram between the used time and the remaining time in the embodiment of the method for estimating the remaining service time of a lithium battery based on load pattern recognition. DETAILED DESCRIPTION

[0021] The technical solution of the present invention is further described below with specific examples, but the protection scope of the present invention is not limited to the following examples.

[0022] Implementation 1: This method for estimating the remaining service life of a lithium battery based on load pattern recognition includes the following steps:

[0023] S1-Load operation superposition mode recognition: the total voltage V, total current I, and energy capacity Q of the lithium battery are sampled. According to the sampling data, the total voltage V k , total current I k , energy capacity Q k , identifying the load as operating in one of the constant current mode, constant resistance mode, constant power mode, and superposition mode;

[0024] S2-Prediction of the remaining time of lithium battery in superposition mode: Construct an SOC table that depends on voltage and current, and calculate the energy capacity Q after k△t time based on the current energy capacity value Q0, voltage V0, and SOC table k and voltage V k , when k=M, SOC occurs M ≤0, the remaining usage time is M△t.

[0025] Implementation 2: The method for estimating the remaining service life of a lithium battery based on load pattern recognition The load operation superposition pattern recognition is specifically as follows:

[0026] The time step of data sampling is Δt, and m+1 data records are taken at the current moment. The total voltage V k , total current I k , energy capacity Q k (k=0,1,…,m), the virtual resistance is The total load power is

[0027] Calculation sequence I k ,R k ,Wk Mean difference and standard deviation S I =std(I k ),S R =std(R k ),W I =std(W k );

[0028] like The load is identified as constant current mode, and the current is like It is identified as constant power mode, and the power is like The load is identified as constant resistance mode, and the resistance is Here It is a threshold value specified based on the data level and experimental experience. If the load does not belong to one of the three single modes mentioned above, the load is identified as a superposition mode.

[0029] In the superposition mode, assuming the load is current I, resistance R, and power W, the energy consumed in the kth time period is The actual energy consumption is E k =Q k -Q k-1 , the difference between the two is The total sum of squared errors is Obtaining the optimal solution through nonlinear numerical optimization methods Right now The load pattern can thus be viewed as a current resistance and power A superimposed combination of three parts.

[0030] The remaining method steps are as described in Implementation 1 and will not be repeated here.

[0031] Implementation method 3: Figure 1 As shown, the remaining time of lithium battery in the superposition mode is predicted in step S2 of the method for estimating the remaining time of lithium battery based on load pattern recognition as follows:

[0032] Assume the voltage range of lithium battery is [V min ,V max ], the current range is [I min ,I max ]; divide the voltage interval m into equal parts, and the voltage step length is The point is V k =V min +kV h ,k=0,1,…,m; divide the current interval n into equal current steps Point I j =Imin +jI h ,j=0,1,…,n; to construct the SOC table SOC that depends on voltage and current kj =F(V k ,I j ), k=0,1,…,m; j=0,1,…,n;

[0033] The actual energy capacity Q of the known battery c , energy capacity offset Q d , the current energy capacity value is Q0, the voltage is V0; suppose the load mode is constant current I c , constant resistance R c , constant power W c The energy consumed by the load running for Δt is The energy after Δt is Q1=Q0-Q Δ , SOC value is

[0034] If SOC1>0, calculate the total current after Δt From SOC1 and I1, the voltage value V1 after the first Δt moment can be calculated using the SOC table; and by analogy, the battery capacity and voltage Q after the kth Δt moment can be obtained. k ,V k , the energy capacity and voltage after (k+1)Δt can be calculated, specifically: Further calculation and the total current By SOC k+1 ,I k+1 Use the SOC table to calculate the voltage V after (k+1)Δt k+1 ; Until the calculation reaches step M, SOC occurs M ≤0, the estimated remaining time is MΔt.

[0035] The remaining method steps are as described in Implementation 1 and will not be repeated here.

[0036] Example: The remaining service time estimation method of lithium battery based on load pattern recognition is applied to the battery No. 6022 to conduct a service time estimation experiment. The estimation results are as follows: Figure 2 、 3 As shown. Among them, Figure 2 The results of dynamic usage and total usage time estimation are shown. As can be seen from the figure, the total usage time of the battery in this cycle is approximately between 7500-7590 seconds, and the total usage time is approximately between 125-132 minutes. The data reflects that during the usage process of more than 2 hours, the error in the determination of the total usage time is approximately 4 minutes, which reflects the accuracy of this estimation method. Figure 3 The graph shows the relationship between elapsed time and remaining time, showing a nearly linear relationship, demonstrating the stability of this estimation method. Remaining battery life estimates for various environments, including drones, forklifts, and electric vehicles, were compared with actual data, and the errors were all on the order of several minutes, falling within an acceptable range.

[0037] This method for estimating the remaining service life of a lithium battery based on load pattern recognition identifies the load pattern. On the basis of the load pattern determination, it borrows historical data and uses differential calculus and difference methods to solve nonlinear optimization problems, thereby more accurately estimating the remaining operating time of the lithium battery.

[0038] The above description shows the main features, basic principles, and advantages of the present invention. It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments or examples described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential features of the present invention. The above embodiments or examples should therefore be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be encompassed within the present invention. Any reference numerals in the claims should not be construed as limiting the claims to which they relate.

[0039] In addition, it should be understood that although this specification is described in terms of implementation methods, not every implementation method contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

Claims

1. A method for estimating the remaining service life of a lithium battery based on load pattern recognition, characterized by: The method The following steps are included: S1-Load operation superposition mode recognition: the total voltage V, total current I, and energy capacity Q of the lithium battery are sampled. According to the sampling data, the total voltage V k , total current I k , energy capacity Q k , the load is identified as operating in one of the constant current mode, constant resistance mode, constant power mode, and superposition mode; when it is identified as superposition mode, it is determined to be constant current I c , constant resistance R c , constant power W c combination of; S2-Prediction of remaining time of lithium battery in superposition mode: This step is specifically as follows: S21: Constructing a voltage- and current-dependent SOC table; S22: The actual energy capacity Q of the known battery c and energy capacity offset Q d , and obtain the energy capacity value Q0 and voltage V0 at the current moment; S23: Take the time step as Δt, start iterative calculation from the current moment (k=0), and calculate the battery capacity Q after the kth Δt moment in sequence k and voltage V k , until the calculation reaches the Mth step, SOC M ≤0, the estimated remaining usage time is MΔt; the iterative calculation is: calculate the energy capacity after time (k+1)Δt: Calculate the SOC value after (k+1)Δt: Calculate the total current after (k+1)Δt: Using the SOC table, according to SOC k+1 and I k+1 Calculate the voltage V after (k+1)Δt k+1 .

2. The method for estimating the remaining service life of a lithium battery based on load pattern recognition according to claim 1, characterized in that: The load operation superposition mode identification described in S1 is specifically as follows: The time step of data sampling is Δt, and m+1 data records are taken, including the total voltage V k , total current I k (k=0,1,…,m); Calculate the current sequence I k , virtual resistance sequence and power sequence The standard deviation S I ,S R ,S W ; Compare the standard deviation with the preset threshold to identify the load as one of the constant current, constant resistance, and constant power modes. If none of them are satisfied, it is identified as the superposition mode. When it is identified as the superposition mode, the total error sum of squares is calculated. The minimum value of

Images