TECHNICAL FIELD
The present disclosure relates to technology for abnormal voltage diagnosis of a
battery.
10 The present application claims the benefit of Korean Patent Application No. 10-
2020-0163366 filed on November 27, 2020 with the Korean Intellectual Property Office,
the disclosure of which is incorporated herein by reference in its entirety.
BACKGROUND ART
15 Recently, there has been a rapid increase in the demand for portable electronic
products such as laptop computers, video cameras and mobile phones, and with the
extensive development of electric vehicles, accumulators for energy storage, robots and
satellites, many studies are being made on high performance batteries that can be recharged
repeatedly.
20 Currently, commercially available batteries include nickel-cadmium batteries,
nickel-hydrogen batteries, nickel-zinc batteries, lithium batteries and the like, and among
them, lithium batteries have little or no memory effect, and thus they are gaining more
attention than nickel-based batteries for their advantages that recharging can be done
3
whenever it is convenient, the self-discharge rate is very low and the energy density is high.
Recently, with the widespread of applications (for example, energy storage
systems, electric vehicles) requiring high voltage, there is a rising need for accurate
diagnosis of abnormal voltage of each of a plurality of battery cells connected in series in a
5 battery pack.
Abnormal voltage conditions of a battery cell refers to faulty conditions caused by
abnormal drop and/or rise of cell voltage due to an internal short circuit, an external short
circuit, a defect in a voltage sensing line, bad connection with a charge/discharge line, and
the like.
10 Attempts have been made to carry out abnormal voltage diagnosis of each battery
cell by comparing a voltage across each battery cell, namely, a cell voltage, at a specific
time with an average cell voltage of the plurality of battery cells at the same time as the
specific time. However, the cell voltage of each battery cell relies on the temperature,
current and/or State Of Health (SOH) of the corresponding battery cell, so it is difficult to
15 achieve accurate diagnosis of abnormal voltage of each battery cell by simply comparing
the cell voltages of the plurality of battery cells measured at the specific time. For
example, when there is a large difference in temperature or SOH between a battery cell
having no abnormal voltage and the remaining battery cells, a difference between the cell
voltage of the corresponding battery cell and the average cell voltage may be also large.
20 To solve this problem, in addition to the cell voltage of each battery cell,
additional parameters such as the charge/discharge current, the temperature of each battery
cell and/or State Of Charge (SOC) of each battery cell may be used for abnormal voltage
diagnosis of each battery cell. However, the diagnosis method using the additional
4
parameters involves a process of detecting each parameter and a process of comparing the
parameters, and thus requires more complexity and a longer time than diagnosis methods
using the cell voltage as a single parameter

WHAT IS CLAIMED IS:
1. A battery diagnosis apparatus for a cell group including a plurality of
battery cells connected in series, the battery diagnosis apparatus comprising:
5 a voltage sensing circuit configured to periodically generate a voltage signal
indicating a cell voltage of each battery cell; and
a control circuit configured to generate time series data indicating a change in cell
voltage of each battery cell over time based on the voltage signal,
wherein the control circuit is configured to:
10 (i) determine a first average cell voltage and a second average cell voltage of each
battery cell based on the time series data, wherein the first average cell voltage is a short
term moving average, and the second average cell voltage is a long term moving average,
and
(ii) detect an abnormal voltage of each battery cell based on a difference between
15 the first average cell voltage and the second average cell voltage.
2. The battery diagnosis apparatus according to claim 1, wherein the control
circuit is configured to:
determine a short/long term average difference corresponding to the difference
20 between the first average cell voltage and the second average cell voltage for each battery
cell,
determine a cell diagnosis deviation corresponding to a deviation between an
average value of short/long term average differences of all the battery cells and the
45
short/long term average difference of the battery cell for each battery cell, and
detect the battery cell which meets a requirement that the cell diagnosis deviation
exceeds a diagnosis threshold as an abnormal voltage cell.
5 3. The battery diagnosis apparatus according to claim 2, wherein the control
circuit is configured to, for each battery cell, generate time series data of the cell diagnosis
deviation, and detect the abnormal voltage of the battery cell from a period of time during
which the cell diagnosis deviation exceeds the diagnosis threshold or the number of data of
the cell diagnosis deviation exceeding the diagnosis threshold.
10
4. The battery diagnosis apparatus according to claim 1, wherein the control
circuit is configured to:
determine a short/long term average difference corresponding to the difference
between the first average cell voltage and the second average cell voltage for each battery
15 cell,
determine a cell diagnosis deviation by calculating a deviation between an average
value of short/long term average differences of all the battery cells and the short/long term
average difference of the battery cell for each battery cell,
determine a statistical adaptive threshold which relies on a standard deviation for
20 the cell diagnosis deviation of all the battery cells,
generate time series data of a filter diagnosis value by filtering time series data for
the cell diagnosis deviation of each battery cell on the basis of the statistical adaptive
threshold, and
46
detect the abnormal voltage of the battery cell from a period of time during which
the filter diagnosis value exceeds a diagnosis threshold or the number of data of the filter
diagnosis value exceeding the diagnosis threshold.
5 5. The battery diagnosis apparatus according to claim 1, wherein the control
circuit is configured to:
determine a short/long term average difference corresponding to the difference
between the first average cell voltage and the second average cell voltage for each battery
cell,
10 determine a normalization value of the short/long term average difference as a
normalized cell diagnosis deviation for each battery cell, and determine a statistical
adaptive threshold which relies on a standard deviation for the normalized cell diagnosis
deviation of all the battery cells,
generate time series data of a filter diagnosis value by filtering time series data for
15 the normalized cell diagnosis deviation of each battery cell on the basis of the statistical
adaptive threshold, and
detect the abnormal voltage of the battery cell from a period of time during which
the filter diagnosis value exceeds a diagnosis threshold or the number of data of the filter
diagnosis value exceeding the diagnosis threshold.
20
6. The battery diagnosis apparatus according to claim 5, wherein the control
circuit is configured to normalize the short/long term average difference by dividing the
short/long term average difference by an average value of short/long term average
47
differences of all the battery cells for each battery cell.
7. The battery diagnosis apparatus according to claim 5, wherein the control
circuit is configured to normalize the short/long term average difference through log
5 calculation of the short/long term average difference for each battery cell.
8. The battery diagnosis apparatus according to claim 1, wherein the control
circuit is configured to generate time series data indicating a change in cell voltage of each
battery cell over time using a voltage difference between a cell voltage average value of all
10 the battery cells and a cell voltage of each battery cell, measured at each unit time.
9. The battery diagnosis apparatus according to claim 1, wherein the control
circuit is configured to:
determine a short/long term average difference corresponding to the difference
15 between the first average cell voltage and the second average cell voltage for each battery
cell,
determine a normalization value of the short/long term average difference as a
normalized cell diagnosis deviation for each battery cell, and generate time series data of
the normalized cell diagnosis deviation for each battery cell,
20 generate the time series data of the normalized cell diagnosis deviation for each
battery cell by recursively repeating the following (i) to (iv) at least once:
(i) determining a first moving average and a second moving average for the time
series data of the normalized cell diagnosis deviation of each battery cell, wherein the first
48
moving average is a short term moving average and the second moving average is a long
term moving average, (ii) determining the short/long term average difference
corresponding to a difference between the first moving average and the second moving
average for each battery cell, (iii) determining the normalization value of the short/long
5 term average difference as the normalized cell diagnosis deviation for each battery cell,
and (iv) generating the time series data of the normalized cell diagnosis deviation for each
battery cell,
determine a statistical adaptive threshold which relies on a standard deviation for
the normalized cell diagnosis deviation of all the battery cells,
10 generate time series data of a filter diagnosis value by filtering the time series data
for the normalized cell diagnosis deviation of each battery cell on the basis of the statistical
adaptive threshold, and
detect the abnormal voltage of the battery cell from a period of time during which
the filter diagnosis value exceeds a diagnosis threshold or the number of data of the filter
15 diagnosis value exceeding the diagnosis threshold.
10. A battery pack comprising the battery diagnosis apparatus according to any
one of claims 1 to 9.
20 11. A vehicle comprising the battery pack according to claim 10.
12. A battery diagnosis method for a cell group including a plurality of battery
cells connected in series, the battery diagnosis method comprising:
49
(a) periodically generating time series data indicating a change in cell voltage of
each battery cell over time;
(b) determining a first average cell voltage and a second average cell voltage of
each battery cell based on the time series data, wherein the first average cell voltage is a
5 short term moving average, and the second average cell voltage is a long term moving
average; and
(c) detecting an abnormal voltage of each battery cell based on a difference
between the first average cell voltage and the second average cell voltage.
10 13. The battery diagnosis method according to claim 12, wherein the step (c)
comprises:
(c1) determining a short/long term average difference corresponding to the
difference between the first average cell voltage and the second average cell voltage for
each battery cell;
15 (c2) determining a cell diagnosis deviation corresponding to a deviation between
an average value of short/long term average differences of all the battery cells and the
short/long term average difference of the battery cell for each battery cell; and
(c3) detecting the battery cell which meets a requirement that the cell diagnosis
deviation exceeds a diagnosis threshold as an abnormal voltage cell.
20
14. The battery diagnosis method according to claim 13, wherein the step (c)
comprises:
(c1) generating time series data of the cell diagnosis deviation for each battery
50
cell; and
(c2) detecting the abnormal voltage of the battery cell from a period of time during
which the cell diagnosis deviation exceeds the diagnosis threshold or the number of data of
the cell diagnosis deviation exceeding the diagnosis threshold.
5
15. The battery diagnosis method according to claim 12, wherein the step (c)
comprises:
(c1) determining a short/long term average difference corresponding to the
difference between the first average cell voltage and the second average cell voltage for
10 each battery cell;
(c2) determining a cell diagnosis deviation by calculating a deviation between an
average value of short/long term average differences of all the battery cells and the
short/long term average difference of the battery cell for each battery cell;
(c3) determining a statistical adaptive threshold which relies on a standard
15 deviation for the cell diagnosis deviation of all the battery cells;
(c4) generating time series data of a filter diagnosis value for each battery cell by
filtering time series data for the cell diagnosis deviation of each battery cell on the basis of
the statistical adaptive threshold; and
(c5) detecting the abnormal voltage of the battery cell from a period of time during
20 which the filter diagnosis value exceeds a diagnosis threshold or the number of data of the
filter diagnosis value exceeding the diagnosis threshold.
16. The battery diagnosis method according to claim 12, wherein the step (c)
51
comprises:
(c1) determining a short/long term average difference corresponding to the
difference between the first average cell voltage and the second average cell voltage for
each battery cell;
5 (c2) determining a normalization value of the short/long term average difference
as a normalized cell diagnosis deviation;
(c3) determining a statistical adaptive threshold which relies on a standard
deviation for the normalized cell diagnosis deviation of all the battery cells;
(c4) generating time series data of a filter diagnosis value by filtering time series
10 data for the normalized cell diagnosis deviation of each battery cell on the basis of the
statistical adaptive threshold; and
(c5) detecting the abnormal voltage of the battery cell from a period of time during
which the filter diagnosis value exceeds a diagnosis threshold or the number of data of the
filter diagnosis value exceeding the diagnosis threshold.
15
17. The battery diagnosis method according to claim 16, wherein the step (c2)
comprises normalizing the short/long term average difference by dividing the short/long
term average difference by an average value of short/long term average differences of all
the battery cells for each battery cell.
20
18. The battery diagnosis method according to claim 16, wherein the step (c2)
comprises normalizing the short/long term average difference through log calculation of
the short/long term average difference for each battery cell.
52
19. The battery diagnosis method according to claim 12, wherein the step (a)
comprises generating time series data indicating a change in cell voltage of each battery
cell over time using a voltage difference between a cell voltage average value of all the
5 battery cells and a cell voltage of each battery cell, measured at each unit time.
20. The battery diagnosis method according to claim 1, wherein the step (c)
comprises:
(c1) determining a short/long term average difference corresponding to the
10 difference between the first average cell voltage and the second average cell voltage for
each battery cell;
(c2) determining a normalization value of the short/long term average difference
as a normalized cell diagnosis deviation for each battery cell;
(c3) generating time series data of the normalized cell diagnosis deviation for each
15 battery cell;
(c4) generating the time series data of the normalized cell diagnosis deviation for
each battery cell by recursively repeating the following (i) to (iv) at least once:
(i) determining a first moving average and a second moving average for the time
series data of the normalized cell diagnosis deviation of each battery cell, wherein the first
20 moving average is a short term moving average and the second moving average is a long
term moving average, (ii) determining the short/long term average difference
corresponding to a difference between the first moving average and the second moving
average for each battery cell, (iii) determining the normalization value of the short/long
53
term average difference as the normalized cell diagnosis deviation for each battery cell,
and (iv) generating the time series data of the normalized cell diagnosis deviation for each
battery cell,
(c5) determining a statistical adaptive threshold which relies on a standard
5 deviation for the normalized cell diagnosis deviation of all the battery cells;
(c6) generating time series data of a filter diagnosis value by filtering the time
series data for the normalized cell diagnosis deviation of each battery cell on the basis of
the statistical adaptive threshold; and
(c7) detecting the abnormal voltage of the battery cell from a period of time during
10 which the filter diagnosis value exceeds a diagnosis threshold or the number of data of the
filter diagnosis value exceeding the diagnosis threshold.