TWI851364B - Outlier diagnosis method for series-connected cells - Google Patents

Abstract

The invention discloses a method for diagnosing cell outliers among series-connected lithium battery, the average voltage changes at the initial stage of discharge Vx and the average voltage change at the end of charge Vy are detected and calculated for each cell, then an outlier index and z-score are calculated, and the z-score is used to find the cell outliers that need attention.

Description

Series Cell Outlier Diagnosis Method

The present invention relates to a method for diagnosing outliers in series-connected cells, and in particular to a method for automatically analyzing and detecting outliers in series-connected lithium batteries.

Previous technologies relied on manual interpretation of individual cell data or graphs, and then focused on outlier cells. Previous technologies could not automate analysis and required a lot of time and manpower.

The first idea of this case is to automate the analysis. Simply input various parameters and the system will automatically provide a list of outlier cells that need attention.

FIG1 is a flow chart of the present invention, which includes the following steps: Step A calculates the average voltage change Vx at the initial stage of discharge of the battery cell; Step B calculates the average voltage change Vy at the end of charge of the battery cell; Step C calculates the outlier index and z-score of the battery cell; Step D sets the threshold value of the outlier (z-score); and Step E outputs the list of batteries that exceed the threshold value. FIG2 will explain the calculation of the average voltage change Vx at the initial stage of discharge, FIG3 will explain the calculation of the average voltage change Vy at the end of charge, and FIG4 will explain the calculation of the outlier index x and z-score of each battery cell.

The reference voltage value Vo in the present invention refers to: the data of the last current before the start of discharge is 0A, that is, the last current before discharge is the reference voltage value Vo for calculating the average voltage change Vx at the initial stage of discharge. Generally, it is the voltage value when fully charged and not yet discharged, which can be the voltage value when floating or static.

Calculation of the average voltage change Vx at the initial stage of discharge: After the start of discharge, the voltage value is sampled n times continuously until the discharge exceeds 5% of the capacity. Each data is compared with the reference voltage value Vo, and the voltage difference is obtained by subtracting the sampled voltage value Vi from the reference voltage value Vo. Then, each voltage difference is added and averaged to obtain the average voltage change Vx at the initial stage of discharge.

The calculation formula of the average voltage change Vx at the initial stage of discharge is: Vx = ; Among them, Vo is the reference voltage value, Vi is the sampling voltage, and n is the number of voltage samples. Floating charge means: the charger continuously charges the fully charged battery with a fixed voltage to maintain its state. At this time, the charging current will be slightly greater than 0A.

Static means: there is no current charging or discharging the battery.

The z-score is a statistical indicator used to calculate the number of standard deviations between a value and the average. The screening criteria can be set to z-score z > 3~6. For example, in a system consisting of 160 cells connected in series, z > 6 can be used as the screening criteria. For the sake of convenience, we have selected the data of 16 cells, as shown in Table 1, to illustrate the spirit of this case. When the number of cells is small, the mean and standard deviation will change. Therefore, z > 3 is used as the screening criteria here. Cell 1 has z > 3 and is considered an outlier and is screened out. The remaining cells have z < 3 and are considered to be cells that are consistent with the overall performance.

Standard deviation calculation formula: σ = ;in,

x is the outlier indicator, x-bar is the average value of x, and Nc is the number of cells;

The calculation formula for the z-score of a single cell is: z = ;in,

x is the cell outlier indicator, x-bar is the mean value of x, and σ is the standard deviation.

Table 1: Data of 16 cells are selected to illustrate the spirit of this case Cell number Average voltage change at the initial stage of discharge (Vx) Average voltage change at the end of charging (Vy) Outlier index (x) (Vx+Vy)/2 z - score 1 409.5444444 89.53846154 249.541453 3.724441209 2 363.1833333 6.230769231 184.7070513 0.276102884 3 347.3055556 5.307692308 176.3066239 -0.17068948 4 351.4888889 5.076923077 178.282906 -0.065577241 5 345.7055556 4.384615385 175.0450855 -0.237786751 6 332.2388889 4.230769231 168.2348291 -0.600002913 7 343.5166667 3.615384615 173.5660256 -0.316453301 8 341.2277778 2.923076923 172.0754274 -0.395733545 9 355.2777778 5 180.1388889 0.033136665 10 352.0277778 5.615384615 178.8215812 -0.036926797 11 361.9444444 5.461538462 183.7029915 0.222700093 12 339.6277778 5 172.3138889 -0.383050524 13 338.7444444 4.461538462 171.6029915 -0.420860928 14 324.9444444 5.230769231 165.0876068 -0.767393785 15 336.95 3.307692308 170.1288462 -0.499266089 16 342.4722222 2.923076923 172.6976496 -0.362639498 x- average (x-bar) 179.5158654 Standard deviation (σ) 18.80163592

Table 1 shows the test results of battery 1:

Average voltage change at the initial stage of discharge Vx = 409.54…;

The average voltage change at the end of charging is Vy = 89.53…;

Outlier index x = (409.54+89.53)/2 = 249.54…;

z = (249.54-179.51)/18.80=3.72….

The data of other cells are expressed in a similar way. Please refer to Table 1. I will not elaborate on them here. In the above table, the average value of the outlier index is 179.51…, and the standard deviation of the outlier index x is 18.80…. Assuming that z score z ＞ 3 is used as the abnormal standard, cell 1 with z ＞ 3 is considered an outlier and will be marked as abnormal, and the rest of the cells with z ＜ 3 are considered to be normal quality cells.

Figure 2 shows the calculation of the average voltage change Vx at the initial stage of discharge.

Step 1: Discharge starts. Discharge starts when the first discharge current appears. When 160 cells are connected in series, we set the current lower limit of the series cells to 20A. That is, when the discharge current of the series cells is greater than 20A, it is judged that discharge has started. However, this current lower limit can be freely adjusted according to the person using this indicator.

Step 2: Sample multiple times and calculate the voltage difference (Vo - Vi) at the time of sampling, that is, the reference voltage value Vo minus the voltage value Vi at the time of sampling.

Step 3: Sum the voltage difference. Sum the voltage differences of all previous samples. .

Step 4: Stop when discharge exceeds 5% capacity.

Step 5: Calculate the average voltage change Vx at the initial stage of discharge, that is: the sum of the voltage differences divided by the number of voltage samples Vx = ;

Where: Vo is the last voltage value when the discharge current is 0A in full charge state, Vi is the sampling voltage, and n is the number of voltage samples.

The voltage change of lithium-iron battery charging curve is very small after entering the platform area. Therefore, this case only uses the change data in the early stage of discharge. The early benchmark is usually 5% of the rated capacity, which actually depends on the battery cell of each brand. The sum of the voltage difference is divided by n records to obtain the average voltage change Vx in the early stage of discharge.

Figure 3 shows the calculation of the average voltage change Vy at the end of charging.

Step 6: Charging starts;

Step 7: Sample multiple times and store the voltage values in a queue;

Step 8: When the data represented by the queue exceeds 10% of the charging capacity; and

Step 9: Calculate the average voltage change Vy at the end of charging; that is, add up the difference between each voltage value in the queue and the first voltage value in the queue and divide it by the number of samples.

Figure 4 shows the calculation of outlier index and z-score for each battery cell.

Step 10: Calculate the outlier index and z-score of each cell;

Calculation of outlier index x: x = (Vx*k1 + Vy*k2)/2; k1 and k2 are adjustment coefficients that can be adjusted as needed. In this case, k1=1 and k2=1 are used for calculation.

The formula for calculating standard deviation is: σ = ;in,

x is the outlier index of each battery cell, x-bar is the average value, and Nc is the number of batteries

The calculation formula for the z-score of a single cell is: z = ;in,

x is the cell outlier indicator, x-bar is the mean value of x, and σ is the standard deviation;

Step 11: Set the outlier (z-score) threshold;

Step 12: Output the list of cells that exceed the threshold value.

The parameters used in the above description are only used as examples to help readers understand the spirit of the case, and are not used to limit the scope of rights of the case. The above description reveals the preferred embodiments and design diagrams of the present invention, but the preferred embodiments and design diagrams are only examples and are not used to limit the scope of rights of the present invention. Anyone who implements the present invention with equal technical means or implements it within the scope of rights covered by the following "patent application scope" shall not deviate from the spirit of the present invention and shall be the scope of rights of the applicant.

Vo: reference voltage value

Vi: sampling voltage

n: Number of voltage samples

Nc: Number of cells

z:z score

σ: standard deviation

k1, k2: correction coefficient

x: outlier indicator

x-Bar: the average value of x

Figure 1 is a flow chart of the present invention. Figure 2 is the calculation of the average voltage change at the beginning of discharge. Figure 3 is the calculation of the average voltage change at the end of charge. Figure 4 is the calculation of the outlier index and z-score of each battery cell.

Claims (6)

A method for diagnosing outliers in series cells comprises the following steps: step A calculates an average voltage change Vx at the initial stage of discharge; step B calculates an average voltage change Vy at the end of charge; step C calculates an outlier index x and a z score of each cell; step D sets a threshold value of an outlier (z score); and step E outputs a list of cells exceeding the threshold value; wherein the calculation of the outlier index x refers to x=(Vx*k1+Vy*k2)/2; and the calculation of the z score refers to:

. The method for diagnosing outliers in series cells as described in claim 1, wherein the step A of calculating the average voltage change Vx at the initial stage of discharge further comprises: step 1: when discharge starts, when the discharge current of the series cells is set to be greater than a pre-specified current (ampere) threshold value, it is determined that discharge has started; step 2: sampling multiple times, calculating the voltage difference (Vo-Vi) at the time of sampling, that is, the reference voltage value Vo minus the voltage value Vi at the time of sampling; step 3: summing up the voltage difference, summing up the voltage differences of all previous samples; step 4: stopping when the discharge exceeds P1% capacity; and step 5: calculating the average voltage change Vx at the initial stage of discharge, that is: the sum of the voltage differences divided by the number of voltage samples;

; Wherein, Vo is the reference voltage value, Vi is the sampling voltage, and n is the number of voltage samples. In the method for diagnosing outliers in series cells as described in claim 2, the P1% described in step 4 refers to 3%~7%. The method for diagnosing outliers in series cells as described in claim 1, wherein the step B of calculating the average voltage change Vy at the end of charging further includes: step 6: charging starts; step 7: sampling multiple times and storing the voltage values in a queue; step 8: when the data represented by the queue exceeds P2% of the charging capacity; and step 9: calculating the average voltage change Vy at the end of charging, that is, summing up the differences between each voltage value stored in the queue and the voltage value of the first value in the queue and dividing by the number of samples. In the method for diagnosing outliers in series cells as described in claim 4, the P2% described in step eight refers to 7%~13%. The method for diagnosing outliers in series cells as described in claim 1, wherein the step C of calculating the outlier index x and z score of each cell further comprises: step 10: calculating the outlier index x and z score of each cell; calculation of outlier index x: x=(Vx*k1+Vy*k2)/2; wherein k1 and k2 are adjustment coefficients and can be adjusted as needed; step 11: setting the threshold value of outliers (z score); and step 12: outputting a list of cells exceeding the threshold value.

Images