JP7802553B2 - Battery characteristic analysis device, battery characteristic measurement device, and battery analysis system - Google Patents

Description

The present invention relates to a battery characteristic analysis device, a battery characteristic measurement device, and a battery analysis system, and in particular to a device that analyzes battery characteristics based on the damped oscillatory current flowing through the battery.

Electric vehicles, such as hybrid vehicles and electric vehicles, which run on battery power, are widely used. For electric vehicles, charging devices connected to the power supply network provided by power supply companies are installed at service stations, parking lots, etc. The batteries in electric vehicles are charged by the charging devices.

Factories, offices, event venues, and other locations use battery-powered electric devices such as forklifts and transport vehicles. To enable multiple electric devices to be used in various locations within a factory, office, or other facility, charging systems have been developed in which charging devices are connected to key points in a locally constructed power supply network.

Technology for controlling multiple charging devices connected to a power supply network is known for electric vehicles and charging systems. With this technology, a control device acquires information indicating the battery's state of charge from each charging device, and controls each charging device according to the battery's state of charge in each charging device. Patent Document 1 describes a distributed power supply system that uses this technology. This distributed power supply system includes multiple power conversion units (charging devices) and a control unit (control unit) that controls each power conversion unit. A battery is connected to each power conversion unit. The control unit acquires the battery's SOC (State of Charge) from each of the multiple power conversion units, and controls the charging and discharging of each power conversion unit according to the SOC.

JP 2016-116428 A Japanese Patent Application Laid-Open No. 2018-179652

Generally, battery performance deteriorates with use. Therefore, one possible system is one in which a charging device measures the degree of battery deterioration, and a control device collects information on the degree of battery deterioration measured by each charging device. One way to measure the degree of battery deterioration is to measure the change in impedance in response to a change in frequency, as described in Patent Document 2, for example. However, this requires sweeping the frequency of the signal voltage applied to the battery over a wide range, which can make measurement difficult.

The purpose of this invention is to measure battery performance in a simple manner.

The present invention is characterized in that damped oscillation data on the damped oscillatory current of a battery to be analyzed is obtained from a damped oscillation circuit that damps oscillation in the current flowing through the battery, two or more peak values are obtained for the magnitude of the time waveform of the damped oscillatory current indicated by the damped oscillation data, and a real part of the impedance of the battery is calculated based on a ratio of one of the two or more peak values to the other and a plurality of first constants determined in advance from the characteristics of the damped oscillation circuit, and an output voltage of the battery is calculated based on the real part of the impedance of the battery, one of the two or more peak values, and a plurality of second constants determined in advance from the characteristics of the damped oscillation circuit.

Desirably, the degree of battery deterioration is determined based on the impedance real part, or based on the impedance real part and the output voltage.

Furthermore, a related technology of the present invention includes, for each of a plurality of batteries, a process of acquiring damped oscillation data about a damped oscillatory current of the battery from a damped oscillation circuit that damps oscillation in a current flowing through the battery to be analyzed, calculating an output voltage for each of the plurality of batteries based on a time waveform of the damped oscillation current indicated by the damped oscillation data acquired for each of the plurality of batteries to obtain machine learning data, calculating an output voltage of a battery different from the plurality of batteries, and fitting the output voltage of the battery different from the plurality of batteries to a machine learning model constructed using the machine learning data to obtain a deterioration level of the battery different from the plurality of batteries. Preferably, based on the damped oscillation current indicated by the damped oscillation data acquired for each of the plurality of batteries, calculating an impedance real part in addition to the output voltage of each of the plurality of batteries to obtain the machine learning data, calculating the impedance real part in addition to the output voltage of the battery different from the plurality of batteries, and fitting the impedance real part and the output voltage of the battery different from the plurality of batteries to a machine learning model constructed using the machine learning data to obtain a deterioration level of the battery different from the plurality of batteries. Preferably, the output voltage of each of the plurality of batteries is a voltage that changes based on changing the load current of each of the plurality of batteries, and the output voltage of a battery different from the plurality of batteries is a voltage that changes based on changing the load current of a battery different from the plurality of batteries.

Preferably, the output voltage of each of the plurality of batteries and the output voltage of a battery different from the plurality of batteries is a relaxation voltage.

Furthermore, a related technology of the present invention is a battery characteristic analysis device that acquires damped oscillation data about a damped oscillatory current of a battery to be analyzed from a damped oscillation circuit that damps oscillation in the current flowing through the battery, and determines a real part of the impedance of the battery based on the damped oscillation data, or determines the real part of the impedance and output voltage of the battery. The battery characteristic measurement device, together with the battery characteristic analysis device, constitutes a battery characteristic analysis system, and includes the damped oscillation circuit and a control unit that generates the damped oscillation data, and the control unit extracts values corresponding to the damped oscillation current flowing through the battery at a plurality of different times, and generates the damped oscillation data based on the values corresponding to the damped oscillation current extracted at the plurality of different times .

Preferably, the damped oscillation circuit includes a primary inductor, a capacitive element, and a switching element, and the primary inductor, capacitive element, and switching element are provided in a path connecting the positive and negative electrodes of the battery. When the switching element conducts in a pulsed manner, the current flowing through the battery undergoes damped oscillation, and the control unit generates the damped oscillation data based on the voltage appearing in a secondary inductor coupled to the primary inductor.

Furthermore, a related technology of the present invention is a battery characteristic analysis device that acquires damped oscillation data about a damped oscillatory current of a battery to be analyzed from a damped oscillation circuit that damps oscillation in the current flowing through the battery, and determines the real part of the impedance of the battery based on the damped oscillation data, or determines the real part of the impedance and output voltage of the battery. This battery characteristic measurement device, together with the battery characteristic analysis device, constitutes a battery characteristic analysis system, and includes the damped oscillation circuit and a control unit that generates the damped oscillation data. The control unit causes the damped oscillatory current to flow through the battery a plurality of times, measures the time waveforms of the damped oscillatory current that flow a plurality of times, extracts values corresponding to the damped oscillatory current at different timings for each measurement, and generates circuit analysis information for analyzing the characteristics of the damped oscillation circuit based on the values extracted for each measurement .

Preferably, the damped oscillation circuit includes a primary inductor, a capacitive element, and a switching element, and the primary inductor, capacitive element, and switching element are provided in a path connecting the positive and negative electrodes of the battery. When the switching element conducts in a pulsed manner, the current flowing through the battery undergoes damped oscillation, and the control unit extracts a value corresponding to the damped oscillation current from the voltage appearing in a secondary inductor coupled to the primary inductor.

A related technology of the present invention is a battery analysis system including a plurality of the battery characteristic measuring devices, wherein the damped oscillation circuit is connected to each of the plurality of batteries connected in series, and each of the damped oscillation circuits includes a resistive element connected in parallel to the capacitive element, and the battery analysis system controls the discharge state of the battery relative to the capacitive element to equalize the charge states of the plurality of batteries .

The present invention allows battery performance to be measured in a simple manner.

FIG. 1 is a diagram illustrating a configuration of a battery analysis system. FIG. 2 is a diagram showing the configuration of the measuring device together with a control unit. 3A and 3B are diagrams showing the time waveforms of each signal, a damped oscillation current, and a damped oscillation voltage. FIG. 2 is a diagram showing each measuring device provided corresponding to each battery together with a control unit. FIG. 4 is a diagram showing a time waveform of a drive signal. FIG. 10 is a diagram showing the configuration of a battery characteristics measuring device according to an application embodiment. FIG. 10 is a diagram showing a state of the battery characteristic measuring device when measuring the resonance frequency of a damped oscillation circuit. FIG. 10 shows the total extraction signal together with the damped oscillatory voltage. FIG. 2 is a diagram illustrating the configuration of a battery degradation analyzer. FIG. 1 is a diagram showing an example of relaxation voltage for a lead battery. 1 is a graph showing the relationship between degradation parameters measured by a reference measuring device and the measured values of the real part of the impedance of a lead-acid battery measured by a battery analysis system. FIG. 2 is a diagram illustrating the configuration of a battery degradation analyzer. FIG. 1 is a diagram showing machine learning data of a lithium-ion battery and a machine learning model of a lithium-ion battery. 10 is a graph showing the change in the real part of the impedance in relation to the degree of deterioration of the battery capacity. FIG. 2 is a diagram conceptually illustrating an example of operation timing of the battery analysis system.

Each embodiment of the present invention will be described with reference to the figures. Identical components shown in multiple figures will be assigned the same reference numerals to simplify their description. Terms indicating directions such as up, down, left, and right in this specification refer to directions in the figures. These terms do not limit the orientation of each component when it is placed.

Figure 1 shows the configuration of a battery analysis system 100 according to an embodiment of the present invention. The battery analysis system 100 includes a battery characteristic measurement device 10, a storage computer 12, and a battery characteristic analysis device 14. The battery characteristic measurement device 10, the storage computer 12, and the battery characteristic analysis device 14 may each be configured as separate computers connected by a communication line. The battery analysis system 100 may also be configured as an integrated unit of hardware. Furthermore, two selected from the battery characteristic measurement device 10, the storage computer 12, and the battery characteristic analysis device 14 may also be configured as an integrated unit of hardware.

A number n of batteries B1 to Bn to be analyzed are connected to the battery characteristic measuring device 10. In this embodiment, the batteries B1 to Bn are connected in series, with both ends of each battery connected to the battery characteristic measuring device 10. In the battery analysis system 100, the battery characteristic measuring device 10 measures the characteristic data required to analyze the characteristics of each battery for each battery and transmits it to the storage computer 12. The storage computer 12 stores the characteristic data. The battery characteristic analysis device 14 reads the characteristic data for each battery from the storage computer 12 and analyzes the characteristics of each battery.

Figure 2 shows the configuration of the measuring device 20, which is the part of the battery characteristic measuring device 10 that acquires characteristic data for one battery Bj (j is an integer between 1 and n), along with the control unit 22. The measuring device 20 includes a damped oscillation circuit 24 and a drive unit 26. The damped oscillation circuit 24 includes a primary inductor 28, a capacitive element 30, a resistive element 32, and a switching element 34.

The primary inductor 28 may be composed of a winding. The primary inductor 28 is coupled to a secondary inductor 38 in the drive unit 26, and together with the secondary inductor 38 forms a transformer. The secondary inductor 38 may also be composed of a winding. The capacitive element 30 may be a capacitor, which is a passive element. The capacitive element 30 may also be an active capacitive element such as a varicap.

One end of the primary inductor 28 is connected to the positive electrode of the battery Bj, and the other end is connected to one end of the capacitive element 30. The other end of the capacitive element 30 is connected to one end of the switching element 34, and the other end of the switching element 34 is connected to the negative electrode of the battery Bj. A resistive element 32 is connected in parallel to the capacitive element 30. A damped oscillatory circuit 24 is formed in the path between the positive and negative electrodes of the battery Bj by the primary inductor 28, capacitive element 30, resistive element 32, and switching element 34. The damped oscillatory circuit 24 includes a resonant circuit containing a resistive component. If the battery Bj being analyzed is a lead battery, the resonant frequency of the damped oscillatory circuit 24 may be 2 kHz to 30 kHz. If the battery Bj being analyzed is a lithium-ion battery, the resonant frequency of the damped oscillatory circuit 24 may be 200 kHz to 30 MHz.

The drive unit 26 includes a drive amplifier 36, a secondary inductor 38, a detection amplifier 40, and peak hold circuits 42-1 to 42-3. The secondary inductor 38 is coupled to the primary inductor 28. One end of the secondary inductor 38 is grounded. The other end of the secondary inductor 38 is connected to the input terminal of the detection amplifier 40. The output terminal of the detection amplifier 40 is connected to the input terminals of the peak hold circuits 42-1 to 42-3, respectively.

The control unit 22 outputs a drive signal P0 to the drive amplifier 36, which then outputs the drive signal P0 to the switching element 34. The drive signal P0 rises from low to high, remains high for a predetermined time, and then changes from high to low. When the drive signal P0 is high, the switching element 34 turns on. The switching element 34 conducts in a pulsed manner in accordance with the drive signal P0. Here, "conducting in a pulsed manner" refers to the switching element 34 changing from off to on, maintaining the on state for a predetermined time, and then changing from on to off.

When the switching element 34 conducts in a pulsed manner, a pulse voltage based on the output voltage of the battery Bj is applied from the positive electrode of the battery Bj to the damped oscillation circuit 24, which is composed of the primary inductor 28, capacitive element 30, resistive element 32, and switching element 34. As a result, a damped oscillatory current flows through the damped oscillation circuit 24, oscillating at the resonant frequency and damping. The damping of the damped oscillatory current is due to losses in the resistive element 32, the resistance component contained in the primary inductor 28, and other factors.

An induced electromotive force having a damped oscillatory waveform is generated in the secondary inductor 38 in response to the damped oscillatory current flowing through the primary inductor 28. Peak hold signals PH1 to PH3 are output from the control unit 22 to the peak hold circuits 42-1 to 42-3, respectively. When, or after, the drive signal P0 rises and the switching element 34 becomes conductive, the peak hold signals PH1 to PH3 rise in steps in the order PH1, PH2, and PH3, and then fall after a predetermined time has elapsed. The peak hold circuits 42-1 to 42-3 extract the value of the induced electromotive force generated in the secondary inductor 38 in response to the peak hold signals PH1 to PH3 in order, and output extraction signals S1 to S3 to the control unit 22, respectively.

The control unit 22 acquires the battery temperature Tmp from the temperature sensor 50 attached to the battery Bj. The control unit 22 generates characteristic data (damped oscillation data) including information indicating the time waveforms of each of the extraction signals S1 to S3, the battery temperature Tmp, a battery ID (IDentification), and the time (timestamp) at which this information was acquired, and transmits this data to the storage computer 12 (Figure 1). The characteristic data is damped oscillation data including information about the damped oscillation current. The storage computer 12 stores the characteristic data transmitted from the control unit 22. Note that the characteristic data may include maximum absolute values Pm1 to Pm3 of each of the extraction signals S1 to S3, instead of or in addition to the information indicating the time waveforms of each of the extraction signals S1 to S3.

Figures 3(a) to (h) show the time waveform of the damped oscillatory current Ia and the time waveforms of each signal used in the battery analysis system 100. Figure 3(a) shows the drive signal P0. Figure 3(b) shows the damped oscillatory current Ia. The drive signal P0 rises from low to high at time t0, remains high for a time Ton from time t0, and then falls from high to low at time t0 + Ton. After time t0, the damped oscillatory current Ia oscillates sinusoidally while attenuating.

Figure 3(c) shows the peak hold signal PH1 together with the damped oscillation voltage Ea, and Figure 3(d) shows the extraction signal S1 together with the damped oscillation voltage Ea. Here, the damped oscillation voltage Ea is the induced electromotive force that appears in the secondary inductor 38 in response to the damped oscillation current Ia. The damped oscillation voltage Ea has a value that corresponds to the damped oscillation current Ia and has the same time waveform as the damped oscillation current Ia. The peak hold signal PH1 rises at time t0 and falls together with the drive signal P0 at time t0 + Ton. While the peak hold signal PH1 remains high, the peak hold circuit 42-1 outputs to the control unit 22 an extraction signal S1 that indicates the maximum value of the damped oscillation voltage Ea from the present time back to time t0, when the peak hold signal PH1 became high.

Figures 3(e) and (g) show peak hold signals PH2 and PH3, respectively, along with the damped oscillation voltage Ea, while Figures 3(f) and (h) show extraction signals S2 and S3, respectively, along with the damped oscillation voltage Ea. Peak hold signals PH2 and PH3 rise at times t2 and t3, respectively, which are times τ2 and τ3 after the rise of drive signal P0, and fall together with drive signal P0 at time t1 + Ton.

While the peak hold signal PH2 remains high, the peak hold circuit 42-2 outputs to the control unit 22 an extraction signal S2 indicating the maximum value of the damped oscillation voltage Ea from the present time back to time t2 when the peak hold signal PH2 became high. While the peak hold signal PH3 remains high, the peak hold circuit 42-3 outputs to the control unit 22 an extraction signal S3 indicating the maximum value of the damped oscillation voltage Ea from the present time back to time t3 when the peak hold signal PH3 became high.

In this way, the control unit 22A extracts a value corresponding to the damped oscillation current Ia (value of the damped oscillation voltage Ea) flowing through the battery Bj at three (multiple) different times to obtain extraction signals S1 to S3, and generates characteristic data (damped oscillation data) based on the extraction signals S1 to S3. The extraction signals S1 to S3 are values corresponding to the damped oscillation current Ia extracted at three (multiple) different times.

The battery characteristic analyzer 14 shown in Figure 1 reads the above characteristic data from the storage computer 12. The battery characteristic analyzer 14 calculates the output voltage Vb and impedance real part Rhf of the battery Bj based on the characteristic data using the following process. Here, the impedance real part refers to the real part when the internal impedance of the battery at a certain frequency is expressed as a complex number.

The battery characteristic analyzer 14 calculates the real part of the impedance Rhf according to Equation 1 based on the maximum absolute value Pm3 of the extraction signal S3 and the maximum absolute value Pm2 of the extraction signal S2.

(Math. 1) Rhf=ln(Pm2/Pm3)・ar-br

Here, ln represents the natural logarithm. The constants ar and br are predetermined based on the characteristics of the damped oscillation circuit 24. (Equation 1) is derived from the fact that the time constants by which the peaks of the damped oscillation current Ia and damped oscillation voltage Ea decay over time are determined according to the real part of the impedance of the battery Bj.

The battery characteristic analyzer 14 calculates the output voltage Vb of battery Bj according to Equation 2.

(Math. 2) Vb=(1/√δ)・k・N・Pm1, δ=exp(−Rhf・tr/(2Lr))

Here, Pm1 is the maximum absolute value of the extraction signal S1. k is the gain of the detection amplifier 40. N is the turns ratio of the windings constituting the secondary inductor 38 to the windings constituting the primary inductor 28. tr is the time from time t0 until the absolute value of the extraction signal S1 reaches its maximum. Lr is the self-inductance of the primary inductor.

The battery characteristic measuring device 10 sequentially acquires characteristic data over time and sequentially transmits the characteristic data to the storage computer 12 over time. The battery characteristic analyzing device 14 sequentially determines the real part of impedance Rhf and the output voltage Vb of the battery Bj over time based on the characteristic data stored in the storage computer 12 over time. The battery characteristic analyzing device 14 may store the real part of impedance Rhf and the output voltage Vb of the battery Bj in association with time, or may store the real part of impedance Rhf and the output voltage Vb of the battery Bj in association with time in the storage computer 12.

The battery characteristic analyzer 14 may determine the degree of deterioration of the battery Bj based on the impedance real part Rhf and the output voltage Vb, using the processing described below.

In the battery analysis system 100 according to this embodiment, if the output voltage Vb of each battery is stored in the storage computer 12, the control unit 22 of the battery characteristic measuring device 10 may execute a process to equalize the output voltages Vb of batteries B1 to Bn. This process controls the discharge state from battery Bj to the capacitive element 30 for battery Bj, which has the largest output voltage Vb among batteries B1 to Bn, thereby equalizing the charge states of batteries B1 to Bn. Figure 4 shows measuring devices 20-1 to 20-n provided corresponding to batteries B1 to Bn, along with the control unit 22. Each of measuring devices 20-1 to 20-n has a configuration similar to that of measuring device 20 shown in Figure 2.

Figure 5 shows the time waveform of the drive signal P0 output from the control unit 22 to the drive amplifier 36 of each of the measuring devices 20-1 to 20-n during operation in the first mode, which acquires characteristic data for each of the batteries B1 to Bn, and operation in the second mode, which equalizes the output voltage Vb of each of the batteries B1 to Bn.

In the first mode of operation, the control unit 22 sequentially outputs three pulses of the drive signal P0 to the measuring devices 20-1 to 20-n, i.e., three pulses for acquiring the extraction signals S1 to S3. Through the above operation, each of the measuring devices 20-1 to 20-n acquires characteristic data. The battery characteristic analyzer 14 calculates the output voltage Vb of each of the batteries B1 to Bn based on the characteristic data.

Figure 5 shows an example in which the output voltage Vb of battery Bj connected to measuring device 20-j is higher than that of the other batteries. In second mode operation, the control unit 22 sets the drive signal P0 output to measuring device 20-j, which is connected to battery Bj with the highest output voltage Vb among batteries B1 to Bn, to high. The control unit 22 also sets the drive signal P0 output to measuring devices other than measuring device 20-j to low. As a result, a damped oscillation current flows through the damped oscillation circuit 24 of battery Bj and measuring device 20-j, after which battery Bj charges capacitive element 30, causing the output voltage Vb of battery Bj to decrease.

The battery analysis system 100 repeatedly operates in the first mode and the second mode. This causes the output voltages Vb of batteries B1 to Bn to converge to the same value, preventing a large electrical load from being placed on any particular battery.

Figure 6 shows the configuration of a battery characteristic measuring device 10A according to an application embodiment of the present invention. In the battery characteristic measuring device 10A, the control unit 22A executes a process for acquiring extraction signals S1 to S3 corresponding to one battery for each of the batteries B1 to B3 in a time-sharing manner.

The battery characteristic measuring device 10A includes measuring devices 20A-1 to 20A-3, each connected to three batteries B1 to B3. Measuring device 20A-1 is similar to measuring device 20-1 shown in Figure 2, except that the peak hold circuits 42-2 and 43-3 have been removed, and instead of grounding the lower end of the secondary inductor 38, it is connected to one end of the secondary inductor 38 of measuring device 20A-2.

Measurement device 20A-2 is configured by removing the detection amplifier 40 and peak hold circuits 42-1 to 42-3 from measurement device 20 shown in FIG. 2, and instead of connecting the upper end of secondary inductor 38 to the lower end of secondary inductor 38 of measurement device 20A-1 and grounding the lower end, it is connected to the upper end of secondary inductor 38 of measurement device 20A-3. Measuring device 20A-3 is configured by removing the detection amplifier 40 and peak hold circuits 42-1 to 42-3 from measurement device 20 shown in FIG. 2, and connecting the upper end of secondary inductor 38 to the lower end of secondary inductor 38 of measurement device 20A-2. In this way, the secondary inductors 38 of measurement devices 20A-1 to 20A-3 are connected in series, the upper end of secondary inductor 38 of measurement device 20A-1 is connected to the detection amplifier 40, and the lower end of secondary inductor 38 of measurement device 20A-3 is grounded.

Cascade-connected delay circuits 52-1 and 52-2 are inserted between the driver amplifier 36 and control unit 22A of measuring device 20A-3, and delay circuit 52-1 is inserted between the driver amplifier 36 and control unit 22A of measuring device 20A-2.

The control unit 22A raises and lowers the drive signal P0 three times at different time periods. The delay circuit 52-1 delays the drive signal P0 by delay time D1 and outputs it to the delay circuit 52-2 and the drive amplifier 36 of the measuring device 20A-2. The delay circuit 52-2 delays the drive signal P0 output from the delay circuit 52-1 by delay time D2 and outputs it to the drive amplifier 36 of the measuring device 20A-3. The delay times D1 and D2 are determined so that the time periods during which the drive signal P0 input to the drive amplifier 36 of each of the measuring devices 20A-1 to 20A-3 is high do not overlap.

When the drive signal P0 input to the drive amplifier 36 of the measuring device 20A-1 rises and falls for the first time, the control unit 22A outputs a pulse of the peak hold signal PH1 to the peak hold circuit 42-1 to acquire the extraction signal S1. As a result, the peak hold circuit 42-1 outputs the extraction signal S1 for battery B1 to the control unit 22A. When the drive signal P0 input to the drive amplifier 36 of the measuring device 20A-1 rises and falls for the second time, the control unit 22A outputs a pulse of the peak hold signal PH2 to the peak hold circuit 42-1 to acquire the extraction signal S2. As a result, the peak hold circuit 42-1 outputs the extraction signal S2 for battery B1 to the control unit 22A. When the drive signal P0 rises and falls for the third time, the control unit 22A outputs a pulse of the peak hold signal PH3 to the peak hold circuit 42-1 to acquire the extraction signal S3. This causes the peak hold circuit 42-1 to output an extraction signal S3 for battery B1 to the control unit 22A.

For measuring device 20A-2, control unit 22A causes peak hold circuit 42-1 to perform the same operation as for measuring device 20A-1, with a delay of delay D1 relative to measuring device 20A-1. As a result, peak hold circuit 42-1 outputs extraction signals S1 to S3 for battery B2 to control unit 22A. For measuring device 20A-3, control unit 22A performs the same operation as for measuring device 20A-1, with a delay of delay D1 + D2 relative to measuring device 20A-1. As a result, peak hold circuit 42-1 outputs extraction signals S1 to S3 for battery B3 to control unit 22A.

The control unit 22A generates characteristic data for each of the batteries B1 to B3 and transmits the characteristic data acquired for each of the batteries B1 to B3 to the storage computer 12. The battery characteristic analyzer 14 calculates the impedance real part Rhf and the output voltage Vb for each of the batteries B1 to B3 based on the characteristic data acquired for each of the batteries B1 to B3.

Here, a configuration for acquiring characteristic data for three batteries B1 to B3 has been described. The battery characteristic measuring device 10A can be modified to acquire characteristic data for two or four or more batteries. In this case, a measuring device similar to measuring device 20A-2 is provided for each battery, and a delay circuit that delays the drive signal P0 is provided between the preceding measuring device and the following measuring device.

Figure 7 shows the state of the battery characteristic measuring device 10B when measuring the resonant frequency of the damped oscillation circuit 24 of the measuring device 20B. The measuring device 20B has a configuration similar to that of the measuring device 20A-1 shown in Figure 6. A reference battery BS is connected to the battery characteristic measuring device 10B instead of the battery Bj. The reference battery BS illustrated in Figure 7 includes a battery B0, a reference capacitor C0, a reference inductor L0, and a reference resistor R0. One end of the reference inductor L0 is connected to the positive terminal of the battery B0. The reference capacitor C0 and the reference resistor R0 are connected in series. One end of the reference capacitor C0 opposite the reference resistor R0 is connected to the other end of the reference inductor L0, and the terminal of the reference resistor R0 opposite the reference capacitor C0 is connected to the negative terminal of the battery B0. The element constants of the reference capacitor C0, the reference inductor L0, and the reference resistor R0 are known. The element constants of the reference capacitor C0, reference inductor L0, and reference resistor R0 may be stored in the storage computer 12, and the control unit 22B or the battery characteristic analyzer 14 may read these element constants.

The control unit 22B repeatedly outputs pulses of the drive signal P0 to the drive amplifier 36 at predetermined time intervals. The control unit 22B repeatedly outputs pulses of the peak hold signal PH to the peak hold circuit 42-1 in synchronization with the drive signal P0.

The control unit 22B delays the rising edge of the pulse of the peak hold signal PH output to the peak hold circuit 42-1 depending on the number of times the pulse of the peak hold signal PH is output. That is, when the control unit 22B outputs the pth pulse of the peak hold signal PH to the peak hold circuit 42-1, the control unit 22B outputs a pulse of the peak hold signal PH whose rising edge is delayed by p·ΔT relative to the rising edge of the drive signal P0 to the peak hold circuit 42-1. The falling edge of the pulse of the peak hold signal PH may be simultaneous with the falling edge of the pulse of the drive signal P0 that is synchronized with it. The peak hold circuit 42-1 sequentially outputs the extraction signal S to the control unit 22B in accordance with the pulse of the peak hold signal PH repeatedly output from the control unit 22B.

According to this processing, as pulses of the drive signal P0 are repeatedly output from the control unit 22B, the switching element 34 conducts in a repeated pulsed manner, and the damped oscillation current Ia repeatedly flows through the damped oscillation circuit 24. The extraction signal S is acquired so that the timing at which the extraction signal S is acquired is delayed by a time ΔT each time the damped oscillation current flows through the damped oscillation circuit 24.

Figure 8 shows the total extraction signal ΣS along with the damped oscillation voltage Ea. The total extraction signal ΣS is a series of extraction signals S repeatedly acquired while delaying the rising edge of the pulse of the peak hold signal PH by a time ΔT, and these signals are arranged on the time axis. The total extraction signal ΣS represents a time waveform that maintains the peak value of the damped oscillation voltage Ea. The control unit 22B determines the frequency of the damped oscillation voltage Ea based on the total extraction signal ΣS. If the element constants of the reference battery BS and the capacitance of the capacitive element 30 included in the measuring device 20B are known, the control unit 22B may determine the inductance value of the primary inductor 28 based on the total extraction signal ΣS. Alternatively, the battery characteristic analysis device 14 may acquire the total extraction signal ΣS via the storage computer 12 and determine the frequency of the damped oscillation voltage Ea or the inductance value of the primary inductor 28. The values determined by measuring device 20B may be used as data for analyzing the degree of deterioration of battery Bj when the battery Bj to be analyzed is connected instead of reference battery BS. In addition, the inductance value of primary inductor 28 determined by measuring device 20B may be used as Lr in Equation 2.

As such, in the battery characteristic measuring device 10B shown in FIG. 7, the control unit 22B conducts the switching element 34 multiple times, and each time the switching element 34 conducts, an extraction signal S is acquired at a different position on the time axis for the damped oscillation current flowing through the damped oscillation circuit 24. That is, the control unit 22B conducts the switching element 34 multiple times, causing a damped oscillation current to flow through the reference battery BS (battery) multiple times. The control unit 22B measures each time waveform of the damped oscillation current flowing multiple times and extracts a value corresponding to the damped oscillation current at a different timing for each measurement as the extraction signal S. The control unit 22B generates a total extraction signal ΣS based on the extraction signal S extracted for each measurement. The total extraction signal ΣS serves as circuit analysis information for analyzing the characteristics of the damped oscillation circuit 24.

The process of acquiring the total extraction signal ΣS and determining the frequency of the damped oscillatory current Ia or the inductance value of the primary inductor 28 may be performed by the battery characteristic measuring device 10A shown in Figure 6. The control unit 22A repeatedly outputs pulses of the drive signal P0 at predetermined time intervals to the drive amplifier 36 of the measuring device 20A-1.

Delay circuits 52-1 and 52-2 delay the pulse of drive signal P0 so that the pulse of drive signal P0 is output in time division from control unit 22A to measuring devices 20A-1, 20A-2, and 20A-3 in that order. Control unit 22A outputs a pulse of peak hold signal PH corresponding to measuring device 22A-1 to peak hold circuit 42-1 in synchronization with the pulse of drive signal P0 input to driving amplifier 36 of measuring device 20A-1. Control unit 22A also outputs a pulse of peak hold signal PH corresponding to measuring device 22A-2 to peak hold circuit 42-1 in synchronization with the pulse of drive signal P0 input to driving amplifier 36 of measuring device 20A-2. Control unit 22A also outputs a pulse of peak hold signal PH corresponding to measuring device 22A-3 to peak hold circuit 42-1 in synchronization with the pulse of drive signal P0 input to driving amplifier 36 of measuring device 20A-3.

For each of measuring devices 20A-1 to 20A-3, control unit 22A delays the rising edge of the pulse of peak hold signal PH output to peak hold circuit 42-1 depending on the number of times the pulse of peak hold signal PH has been output. Through this processing, the following operation is performed for each of measuring devices 22A-1 to 22A-3. That is, as pulses of drive signal P0 are repeatedly output from control unit 22B, switching element 34 repeatedly conducts in a pulsed manner, causing a damped oscillation current to repeatedly flow through damped oscillation circuit 24. The extraction signal S is acquired so that the timing at which the extraction signal S is acquired is delayed by a time ΔT each time a damped oscillation current flows through damped oscillation circuit 24.

Here, we have described the process of acquiring the total extraction signal ΣS for each of the three measuring devices 20A-1 to 20A-3 and determining the frequency of the damped oscillatory current or the inductance value of the primary inductor 28. The number of measuring devices to be processed may be two, or four or more. The battery characteristic measuring device may be configured such that a delay circuit that delays the drive signal P0 is provided between the preceding measuring device and the following measuring device.

Figure 9 shows the configuration of the battery degradation analyzer 60 configured in the battery characteristic analysis device 14. The battery degradation analyzer 60 is used to determine the degree of degradation of a lead-acid battery. The battery degradation analyzer 60 receives the impedance real part Rhf, the battery output voltage Vb, and the battery temperature Tmp, which are acquired sequentially over time. Here, the output voltage Vb is the relaxation voltage that increases over time when the battery changes from a state in which a load current is flowing through it (on state) to a state in which the load current is interrupted (off state). Alternatively, the output voltage Vb is the relaxation voltage that decreases over time when the battery changes from an off state to an on state.

The battery degradation analyzer 60 fits at least one of the impedance real part Rhf, the relaxation voltage of the battery output voltage Vb, and the battery temperature Tmp into a pre-constructed machine learning model to determine the degradation levels SOH_R and SOH_C. SOH_R is defined, for example, as a value indicating the rate (percentage increase) of increase in the impedance real part Rhf compared to a new battery. In this case, the larger the SOH_R, the greater the degradation level of the battery. SOH_C is defined, for example, as a value indicating the rate (percentage decrease) of decrease in discharge capacity (mAh) compared to a new battery. The larger the SOH_C, the greater the degradation level of the battery.

The machine learning model may be constructed, for example, by obtaining machine learning data for multiple batteries with different cumulative values (usage amounts) of charge and discharge charges since the start of use. For example, the battery characteristic analysis device 14 acquires characteristic data for each of the multiple batteries with different usage amounts. Furthermore, the SOH_R and SOH_C for each of the multiple batteries with different usage amounts are measured using a device separate from the battery analysis system 100. The battery characteristic analysis device 14 then matches the characteristic data with the measured SOH_R and SOH_C using a machine learning algorithm to obtain machine learning data. The machine learning data is stored in the storage computer 12 and loaded from the storage computer 12 into the battery degradation analyzer 60.

The upper part of Figure 10 shows an example of the relaxation voltage for a lead-acid battery. The horizontal axis shows time, and the vertical axis shows the battery output voltage. Output voltage 62-1 shows the relaxation voltage for a new battery, and output voltage 62-2 shows the relaxation voltage for a degraded battery with a non-zero usage amount. The battery changes from the off state to the on state at time 40 seconds, and its output voltage decreases. Furthermore, the battery changes from the on state to the off state at time 240 seconds, and its output voltage increases.

The lower part of Figure 10 shows the measurement results of the impedance real part of a lead-acid battery measured by the battery analysis system 100 according to an embodiment of the present invention. The horizontal axis shows time, and the vertical axis shows the impedance real part. The impedance real part 64-1 shows the value of a new battery, and the impedance real part 64-2 shows the value of a deteriorated battery with a non-zero usage amount. The lower part of Figure 10 shows that the fluctuations in the impedance real part are small even when the output voltage fluctuates, as shown in the upper part of Figure 10.

Figure 11 shows a graph correlating degradation parameters measured by a general reference measurement device (reference measurement device) with the measured values of the real part of the impedance of a lead-acid battery measured by the battery analysis system 100 according to this embodiment. The horizontal axis represents the degradation parameters measured by the reference measurement device. The smaller the value of the degradation parameter in Figure 11, the greater the degree of degradation. The vertical axis represents the real part of the impedance measured by the battery analysis system 100. There is a tendency for the real part of the impedance to be larger as the degradation parameter measured by the reference measurement device becomes smaller, i.e., the greater the degree of degradation. A machine learning model for lead-acid batteries may utilize this tendency.

Figure 12 shows the configuration of the battery degradation analyzer 70 configured in the battery characteristic analysis device 14. The battery degradation analyzer 70 is used to determine the degree of degradation of a lithium-ion battery. The real part of the impedance Rhf, the battery output voltage Vb, and the battery temperature Tmp, which are acquired sequentially over time, are input to the battery degradation analyzer 70. The output voltage Vb, which is the relaxation voltage, is input to the battery degradation analyzer 70.

The battery degradation analyzer 70 fits at least one of the impedance real part Rhf, the relaxation voltage of the battery output voltage Vb, and the battery temperature Tmp into a pre-constructed machine learning model to determine the degradation levels SOH_R, SOH_C, and SOH_S. SOH_S indicates, for example, the amount of metal deposition in the active material of the battery. The greater the amount of metal deposition, the greater the degradation level of the battery. SOH_R is defined, for example, as a value indicating the percentage decrease in the impedance real part compared to a new battery. In this case, the greater the SOH_R, the greater the degradation level of the battery. SOH_C is defined, for example, as a value indicating the percentage decrease in discharge capacity (mAh) compared to a new battery. The greater the SOH_C, the greater the degradation level of the battery.

The machine learning model may be constructed, for example, by obtaining machine learning data for multiple batteries with different usage amounts. For example, the battery characteristic analysis device 14 acquires characteristic data for each of the multiple batteries with different usage amounts. Furthermore, the SOH_R, SOH_C, and SOH_S for each of the multiple batteries with different usage amounts are actually measured by a device separate from the battery analysis system 100. The battery characteristic analysis device 14 then associates the characteristic data with the actually measured SOH_R, SOH_C, and SOH_S using a machine learning algorithm to obtain machine learning data. The machine learning data is stored in the storage computer 12 and is read from the storage computer 12 into the battery degradation analyzer 60.

Figure 13 shows example machine learning data for lithium-ion batteries and results obtained from a machine learning model for lithium-ion batteries. The top section of Figure 13 shows a graph illustrating the machine learning data. The graph on the left side of the top section shows machine learning data that correlates output voltage with discharge capacity for a new battery and three degraded batteries A to C with different amounts of usage. The horizontal axis shows discharge capacity, and the vertical axis shows output voltage. The graph on the right side of the top section shows relaxation voltage as machine learning data for a new battery and three degraded batteries A to C with different amounts of usage. The horizontal axis shows time, and the vertical axis shows output voltage.

The lower part of Figure 13 shows an example of results obtained using a machine learning model for a lithium-ion battery. The left side of the lower part of Figure 13 shows a graph (based on a machine learning model) in which a predicted value is obtained by providing an actual measured value of discharge capacity. The horizontal axis shows the actual measured value of discharge capacity, and the vertical axis shows the predicted value. The actual measured value and the predicted value are matched by a straight line on the graph. The right side of the lower part of Figure 13 shows a graph (based on a machine learning model) in which a predicted value is obtained by providing an actual measured value of the real part of impedance for a battery with a different usage amount from degraded batteries A to C. The horizontal axis shows the actual measured value of the real part of impedance, and the vertical axis shows the predicted value. The actual measured value and the predicted value are matched by a straight line on the graph. The battery degradation analyzer 70 may calculate SOC_R and SOH_C based on the machine learning model shown in Figure 13.

Figure 14 shows a graph that correlates changes in the impedance real part with the degree of battery capacity degradation. The impedance real part was measured using a battery analysis system 100 according to an embodiment of the present invention. Batteries for which the impedance real part indicated by the open circles was measured have a larger amount of metal deposition than batteries for which the impedance real part indicated by the filled circles was measured. As shown in Figure 14, batteries with a larger amount of metal deposition have a larger decrease in the impedance real part relative to a decrease in battery capacity. In other words, in lithium-ion batteries, the greater the degree of battery degradation, the stronger the tendency for the impedance real part to decrease. A machine learning model for lithium-ion batteries may utilize this tendency.

Figure 15 conceptually shows an example of the operation timing of the battery analysis system 100. The upper part of Figure 15 shows the load current of multiple series-connected batteries B1 to Bn. The horizontal axis represents time, and the vertical axis represents the load current. A positive load current indicates that power is being output from batteries B1 to Bn, and a negative load current indicates that power is being supplied to batteries B1 to Bn. The lower part of Figure 15 shows the output voltages Vb of two of batteries B1 to Bn. The horizontal axis represents time, and the vertical axis represents the output voltage.

The battery analysis system 100 constantly calculates the real part of the impedance and the output voltage of each battery. The battery analysis system 100 also constantly executes a process to equalize the output voltages of multiple batteries. As shown in the bottom part of Figure 15, the output voltages of the two batteries converge to the same value over time.

In the example shown in the upper part of Figure 15, the load current changes from 0 to a positive value at time T1, and from a positive value to 0 at time T2. That is, each battery changes from an off state to an on state at time T1, and from an on state to an off state at time T2. Therefore, at times T1 and T2, the relaxation voltage of each battery may be measured, and machine learning data may be stored in the storage computer 12.

With the battery analysis system 100, a damped oscillatory current flows through each battery by turning the switching element of the battery characteristic measurement device on and off, and the battery performance is measured based on the damped oscillatory current. Therefore, there is no need to sweep the frequency of the measurement signal, and the performance of each battery can be measured in a simple manner.

10 Battery characteristic measuring device, 12 Storage computer, 14 Battery characteristic analyzing device, 20, 20-1 to 20-n, 20A-1 to 20A-3, 20B Measuring device, 22, 22A, 22B Control unit, 24 Damped oscillation circuit, 26 Drive unit, 28 Primary inductor, 30 Capacitive element, 32 Resistive element, 34 Switching element, 36 Drive amplifier, 38 Secondary inductor, 40 Detection amplifier, 42-1 to 42-3 Peak hold circuit, 50 Temperature sensor, 52-1, 52-2 Delay circuit, 60, 70 Battery degradation analyzer, 62-1, 62-2 Output voltage, 64-1, 64-2 Impedance real part, B1 to Bn Battery, B0 Reference battery, L0 Reference inductor, C0 Reference capacitor, R0 Reference resistive element.

Claims (2)

obtaining damped oscillation data on the damped oscillation current of the battery from a damped oscillation circuit that damps oscillation of the current flowing through the battery to be analyzed;
a battery characteristic analysis device that acquires two or more peak values for the magnitude of a time waveform of a damped oscillation current indicated by the damped oscillation data, calculates a real part of an impedance of the battery based on a ratio of one of the two or more peak values to the other and a plurality of first constants that are predetermined based on characteristics of the damped oscillation circuit, and calculates an output voltage of the battery based on the real part of the impedance of the battery, one of the two or more peak values, and a plurality of second constants that are predetermined based on characteristics of the damped oscillation circuit. The battery characteristic analyzer according to claim 1,
A battery characteristic analysis device, characterized in that a degree of deterioration of the battery is determined based on the impedance real part, or based on the impedance real part and the output voltage.