JP3528428B2 - Electric vehicle power control device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a device for controlling discharge power and regenerative charge power of a battery mounted on an electric vehicle.

[0002]

2. Description of the Related Art A battery, which is an energy source of an electric vehicle, has a plurality of single cells connected in series and is called an assembled battery. In electric vehicles, in order to manage the operating voltage range of the battery and protect the battery, it is effective to calculate the maximum discharge power and the maximum charge power of the battery pack and control the maximum power during discharge and regenerative charge. is there.

[0003]

However, there is a variation in the terminal voltage of each single cell in the assembled battery in use, and when the maximum charge / discharge power is calculated using the terminal voltage of the assembled battery, the result of the calculation is An error occurs because only the average terminal voltage of a single cell is reflected.

An object of the present invention is to calculate an accurate maximum charge / discharge power of an assembled battery in consideration of variations in the terminal voltage of each single cell, and to accurately control the maximum power during discharge and regenerative charge. Another object of the present invention is to provide a power control device for the electric vehicle.

[0005]

(1) According to the invention of claim 1, voltage measuring means for measuring a terminal voltage of an assembled battery in which a plurality of unit cells are connected in series, and current flowing through the assembled battery are measured. The current measuring means and a plurality of measurement results measured by the voltage measuring means and the current measuring means are linearly regressed on a two-dimensional plane of voltage and current, and the voltage axis intercept of the regression line is the open circuit voltage Eo of the battery pack, the regression. The slope of the straight line is the internal resistance R of the battery pack, and the maximum charge power PC and maximum discharge power PD of the battery pack are calculated based on the open circuit voltage Eo, the internal resistance R, the maximum allowable voltage Vmax of the battery pack and the discharge end voltage Vmin of the battery pack. And a ratio Rmax / Rave between the average value Rave and the maximum value Rmax of the internal resistance of a plurality of single cells as a correction coefficient M, the maximum charging power PC to PC / M, and the maximum discharge power PD to PD / M
And a power correction means for correcting each of them. (2) In the electric power control device for an electric vehicle according to claim 2, the open voltage Eo, the terminal voltage of the assembled battery, and
The correction coefficient M is calculated based on the specified voltage of the single cell and the number of single cells. (3) The electric power control device for an electric vehicle according to a third aspect of the present invention includes alarm means for issuing an alarm when the correction coefficient M exceeds a predetermined value.

[0006]

(1) According to the first and second aspects of the present invention, an accurate maximum charge / discharge power can be calculated, and the upper and lower limit voltages of the assembled battery during discharging and regenerative charging can be accurately calculated. Can be controlled. (2) According to the invention of claim 3, it is possible to accurately detect an abnormality in the battery pack.

[0007]

1 is a block diagram showing the structure of a traveling drive mechanism of an electric vehicle according to an embodiment. During power running, the battery 11 is discharged to supply DC power to the inverter 12, and the inverter 12 converts the DC power into AC power and applies it to the motor 13. As a result, the motor 13 generates running energy and the vehicle runs. On the other hand, during regeneration, the running energy of the vehicle is converted back into electric energy via the motor 13 and the inverter 12, and the battery 1
1 is charged and the vehicle is regeneratively braked. The battery 11 is composed of n single cells 111 to 11n, and the voltage across each single cell is referred to as a cell voltage in this specification. The voltage sensor 14 detects the terminal voltage (hereinafter referred to as the total voltage) Vt of the battery 11, and the current sensor 15 detects the battery 1
The current I flowing through 1 is detected. The current I is positive when flowing from the battery 11 to the inverter 12 when driving the motor, and is negative when flowing from the inverter 12 to the battery 11 during regenerative charging. The battery controller 16 calculates the maximum discharge power and the maximum charge power of the battery 11 based on the voltage V and the current I detected by the voltage sensor 14 and the current sensor 15, and controls the output of the inverter 12 based on the calculation result. And regenerative control. The cell controller 17 controls charge / discharge of each of the unit cells 111 to 11n of the battery 11, and also controls each of the unit cells 111 to 11n.
The terminal voltage (hereinafter, referred to as cell voltage) Vc of is detected.

<< Characteristics of Battery >> Here, the characteristics of the battery will be described. Certain types of batteries, such as lithium-ion batteries and nickel-hydrogen batteries, have the following characteristics. (1) As shown in FIG. 2, the depth of discharge of the battery (hereinafter, D
OD (Depth of Discharge) is low (~ 60)
%), The internal resistances during charging and discharging are almost the same. (2) The linearity of the voltage V-current I characteristic during charge / discharge is good. If such a characteristic of this type of battery is utilized, the DO
It is possible to accurately calculate the maximum discharge power and the maximum charge power according to the state of the battery such as D and temperature. The battery is not limited to the lithium ion battery or the nickel hydrogen battery, and may be any battery having the above characteristics.

Generally, a battery has a characteristic that when discharged at a constant current, the voltage decreases as the discharge reaction progresses, and when charged at a constant current, the voltage rises as the charging reaction progresses. FIG. 3 shows changes in the total voltage Vt of the battery when discharged with a constant current. From time t1 to t3
Up to the total voltage Vt
Changes as shown in the figure. At the discharge start time t1, the total voltage Vt instantly drops from V0 to V1 at 0.1 mS or less. This transient voltage VX is a voltage drop due to the liquid resistance and contact resistance of the battery. Next, in the period T1 from time t1 to time t2, the total voltage Vt rapidly decreases from V1 to V2. This fall time is 100 mS or less, and the transient voltage VY is a voltage drop due to the charge transfer resistance of the battery.
Further, the total voltage Vt gradually decreases from V2 to V3 in the period T2 from time t2 to time t3. This transient voltage VZ is a voltage drop due to concentration polarization of the electrolytic solution, and this region is generally called a diffusion region. After that, when the discharge stop time t3 has passed, the total voltage Vt recovers rapidly. Periods T1 and T3 are transient states that change in a short time, and period T2 is a normal use state of the battery. In this way, the terminal voltage of the battery changes according to the discharge current, and also changes over time of the discharge reaction. Therefore,
When the maximum discharge power of the battery is calculated based on the terminal voltage in the transient state of the discharge reaction, an error may occur, and in some cases, the terminal voltage at the time of discharging may fall below the discharge end voltage of the battery. The discharge end voltage is the allowable minimum voltage of a battery that cannot continue discharging any more.

<< Calculation Method of Maximum Charge / Discharge Power >> Next, a calculation method of the maximum discharge power and the maximum charge power of the battery will be described.
First, the V-I characteristic of the battery being discharged is sampled, and the sampling result is plotted on a V-I graph as shown in FIG. 4 (indicated by X in the figure). As described above, in this type of battery, the internal resistances during charging and discharging are almost the same, and the linearity of the V-I characteristic is good.
The −I characteristic can be linearly regressed, and the regression line can be extended to the charge side and the discharge side. In the figure, the Vt axis intercept Eo of the regression line represents the open circuit voltage of the battery, and the slope of the regression line represents the internal resistance R of the battery. The regression line is

[Equation 1] V = Eo-IR Can be expressed as

The current ICmax at the intersection A between the regression line and the maximum allowable voltage Vmax during charging gives a charge allowable value, and at the intersection A, the following equation is established.

## EQU00002 ## Vmax = Eo-ICmax.R Similarly, the current IDmax at the intersection B between the regression line and the discharge end voltage Vmin at the time of discharge gives a discharge allowable value, and at the intersection B, the following equation is established.

[Equation 3] Vmin = Eo-IDmax · R

The maximum charging power PC can be calculated by the above equation 2

[Equation 4] PC = Vmax ・ ICmax = Vmax ・ (Eo-Vmax) / R The maximum discharge power PD is

## EQU00005 ## PD = Vmin.IDmax = Vmin. (Eo-Vmin) / R. The sampling value of the VI characteristic during discharging is a value according to the battery state such as the DOD and temperature of the battery, and the maximum charging power PC and the maximum discharging power PD obtained by linear regression of such sampling values. Is, of course, electric power according to the state of the battery such as DOD and temperature.

By the way, the discharge end voltage Vmin of a battery mounted on an electric vehicle is usually

[Formula 6] Vmin ≧ Eo / 2. On the other hand, in general, the maximum output P of a battery, that is, the maximum discharge power P is

## EQU7 ## It is known that V = Eo / 2 is obtained. Therefore, the discharge end voltage Vmin of the electric vehicle battery is

When set within the range of Vmin> Eo / 2, the maximum discharge power PD calculated by the above-mentioned method becomes smaller than the maximum discharge power of a general battery. In this specification, the discharge end voltage Vmin is expressed by Equation 6
The discharge power that is set within the range and is calculated for the set discharge end voltage Vmin is the maximum discharge power of the battery.

<< Cell Voltage Variation Correction Coefficient >> FIG. 5 shows the VI characteristics of a single cell of the battery 11. The cell voltages of the n single cells forming the battery 11 have variations due to the deterioration state and the like, and even if the same current flows, the cell voltages do not become the same. The open-circuit voltage Eo of each unit cell can be set to a substantially equal value by equalizing the voltage variations by the cell controller 17 as long as it is within a predetermined discharge amount. The internal resistance of a single cell showing an average voltage drop is R
ave and the internal resistance of the deteriorated cell having the largest voltage drop is Rmax, the cell voltage drops by I1 · Rave from the open voltage Eo in the average single cell with respect to the discharge current I1, and the open voltage Eo in the deteriorated cell. To I1 ・ Rma
The cell voltage drops by x.

As described above, the depth of discharge DOD is within a predetermined value (for example, 60 in the case of the lithium ion battery).
%), The internal resistances during charging and discharging are almost the same, and the linearity of the VI characteristic is good. Therefore, when the DOD is within a predetermined value and discharging, the cell voltage Vc of any one of the n unit cells is a predetermined voltage (for example,
3.4V) or less, the state is a predetermined time (for example,
30 mS) or more, based on the total voltage Vt and the open circuit voltage Eo estimated by the calculation of the charge / discharge power,
The ratio between the average internal resistance of the single cell and the maximum internal resistance of the deteriorated single cell is obtained, and is set as the variation correction coefficient M. As the total voltage Vt, for example, a moving average value before and after 50 mS centered around a time point when any cell voltage Vc drops to a predetermined voltage, for example, 3.4 V or less is used.

## EQU00009 ## M = Rmax / Rave = I1.Rmax / I1.Rave = (Eo-3.4.n) / (Eo-Vt) The correction coefficient M calculated by Equation 9 is obtained in the past. Digital filtering such as weighted average and moving average with M is performed. For example, a weighted average of the previous value and 1/2 may be taken.

<< Sampling Method of Battery VI Characteristics >>
Next, a method of sampling the V-I characteristic of the battery will be described. The relationship between the voltage V and the current I of the battery is expressed by the above mathematical expression 1. However, as described above, the total voltage Vt of the battery changes with the passage of time of the discharge reaction, and even if the voltage is sampled at the same discharge current, the same voltage cannot be obtained if the reaction steps are different. Conversely, even if the current is sampled at the same voltage, the same current cannot be obtained if the reaction steps are different. That is, since the total voltage Vt and the discharge current I change depending on the stage of the chemical reaction of the battery itself, in order to accurately estimate the ability of the battery every moment, the reaction stage of the battery should be taken into consideration in the V-I characteristic. Need to sample.

FIG. 6 shows the total voltage Vt and the discharge current I during discharge.
5 is a diagram illustrating the sampling timing of FIG. Generally, a battery has a property that a change in voltage is delayed with respect to a change in current when the discharge current decreases. Therefore, when sampling the total voltage Vt and the discharge current I, a transient phenomenon between different reaction forms such as discharge reaction to charge reaction, discharge reaction to discharge stop, or equilibrium state to discharge reaction or charge reaction and discharge current decrease. In order to exclude the time, the rise of the discharge current is detected, and the total voltage Vt and the discharge current I when the discharge current increases are sampled. Further, in order to avoid sampling of the unstable total voltage Vt and the discharge current I in the transient region (the period T1 in which the transient voltages VX and VY of FIG. 3 occur), a predetermined time Δ from the rise of the discharge current is avoided.
The total voltage Vt after t and the discharge current I are sampled.

Here, the discharge current rises when the current I and its rate of change dI / dt are both positive. The rising point of this discharge current is the starting point of the discharge reaction of the battery. In the example of FIG. 6, t1, t3, and t5 are the rising points of the discharge current, and a predetermined time Δ from those points.
At t2, t4, and t6 after t, the total voltages V1, V2, V3 and the discharge currents I1, I2, I3 are sampled, respectively. As a result, the measurement in the unstable transient region where the state of the battery changes rapidly can be avoided, and the total voltage Vt and the discharge current I of the battery in the stable diffusion region can be measured. When sampling the V-I characteristic, if there is a new rise of the discharge current within a predetermined time from the rise of the discharge current, timing is restarted from that point, and the total voltage Vt and the discharge are reached after the predetermined time. The current I is sampled.

By the way, if the sampling timing of the VI characteristic is set to only one point after a predetermined time from the rising of the discharge current, the following problems occur. FIG. 7 shows how the discharge current I and the total voltage Vt are sampled after a predetermined time Δt from the rise of the discharge current. Also, FIG.
Is a plot of the sampling results on a VI graph and linear regression. The maximum discharge power PD of this battery is
It is obtained by the above-mentioned formula 5. The calculated maximum discharge power PD is obtained based on the sampling result after the predetermined time Δt from the rising of the discharge current, and is therefore the maximum power that can be discharged for the predetermined time Δt. However, when the discharge exceeding this maximum discharge power PD in Δt time, as indicated by the straight line C in FIG. 8, the total voltage Vt in the middle discharge becomes less discharge end voltage Vmin.

That is, since the dischargeable electric power of the battery varies depending on the reaction stage, it is necessary to sample the VI data according to the required discharge time. Therefore, after the discharge current rises, the VI characteristic is sampled at a plurality of times. FIG. 9 shows how the discharge current I and the total voltage Vt are sampled after a predetermined time Δt1 and Δt2 from the rising of the discharge current. Also, FIG.
0 is the one obtained by plotting the sampling result on the VI graph and performing linear regression. In FIG. 10, the straight line D is a regression line based on the sampling data (x mark) after the predetermined time Δt1 from the rising of the discharge current, and the straight line E is the sampling data (◯ mark) after the predetermined time Δt2 from the rising of the discharge current. It is a regression line based on this. Also, the straight line F
Is a predetermined time Δt1 after the rising of the discharge current and Δt
It is a regression line based on the sampling data after 2 (both x and ◯). With this regression line F, the maximum discharge power PD
Can be obtained, it is possible to obtain the average maximum power for various discharge forms with different discharge currents and discharge times.

FIG. 11 shows an example in which the VI characteristic is sampled at a plurality of points in a normal driving pattern of an electric vehicle. In this example, 1 from the rise of the discharge current
Sampling is performed after 2 seconds and 3 seconds. In addition, FIG. 12 is a graph obtained by plotting the sampling results on a VI graph and performing linear regression. In FIG. 12, the straight line G is the sampling data (marked with x) one second after the rise of the discharge current.
And a straight line H is a regression line based on sampling data (marked with a circle) 3 seconds after the rise of the discharge current. The straight line J is a regression line based on the sampling data (both x and ◯) 1 second and 3 seconds after the rise of the discharge current.

FIG. 13 shows the sampling data Δt and the discharge current I based on the sampling data shown in FIGS.
It is classified by. 5 samples were collected in 1 second sampling and 3 samples in 3 seconds sampling.
Individual data were collected. However, as shown in FIG. 12, the sampling data after 1 second with a small Δt (x mark)
Tends to concentrate in the low current region, and therefore the accuracy of the regression calculation based on these data is low. On the other hand, although the sampling data (marked with a circle) after 3 seconds in which Δt is large is distributed in a wide current range, the number of data tends to be small, and the regression calculation accuracy is also low. However, the sampling data (both × and ◯) at a plurality of time points from the rise of the discharge current, that is, after 1 second and 3 seconds, naturally have a large number of data, and are distributed in a wide range of the current I and the voltage V. Therefore, the accuracy of the regression calculation is increased. Therefore, by calculating the maximum discharge power PD from the regression line J based on these data, it is possible to obtain the ideal average maximum power for various discharge forms with different discharge currents and discharge times. Also, if the accuracy of the regression calculation improves based on the sampling data at multiple time points,
As shown in FIG. 4, it is possible to extend the calculated regression line to the charging side to obtain an accurate maximum charging power PC.

In the sampling examples shown in FIGS. 9 to 12, the sampling is performed at two points after the rise of the discharge current, but the sampling timing may be three or more. Further, in sampling the VI characteristic, if there is a new rise of the discharge current within a predetermined time from the rise of the discharge current, timing is restarted from that time point, and the total voltage Vt and the discharge are reached after the predetermined time. The current I is sampled.

<< Method of Storing Sampling Data of VI Characteristics >> The data sampled at a plurality of timings described above are stocked by the following method. The range of the discharge current I is divided into a plurality of areas, and a predetermined number of stock memories are prepared for each area. For example, as shown in FIG. 14, the discharge current range is divided into five areas, and three stock memories are prepared for each area. Then, during the predetermined sampling time, the current im and the voltage vm at the above-mentioned timings.
(M indicates the sampling order) and
Stock by classifying by current area. When the number of data in the current region reaches a predetermined number, the oldest data is erased and the latest data is stocked. For example, in the example of FIG. 14, when the data (i8, v8) is sampled and the data is included in the I2 to I3 area, the oldest data (i3, v3) in that area is erased and the latest data is used instead. The data (i8, v8) of is stored. According to this sampling data stocking method, since only a predetermined number of data sufficient for linear regression is stocked for each divided current region, a straight line of the VI characteristic based on the sampling data concentrated in a specific divided current region is stocked. Recursion can be avoided, and accurate linear regression can be performed based on sampling data of a wide range of total voltage and discharge current, and accurate maximum charge / discharge power P
D and PC can be estimated. Further, since a predetermined number of sampling data are stocked for each divided current region, accurate linear regression can be performed using sampling data in a wide range within the discharge current range, and accurate maximum charge / discharge power P
In addition to estimating D and PC, there is no need to secure a huge memory capacity in the controller.

The V-I characteristic is sampled within a predetermined time or every predetermined amount of discharged electricity, and the maximum discharge power PD and the maximum charge power PC are calculated based on the data stocked by the above method. Maximum discharge power PD and maximum charge power PC
When the calculation of is finished, all the sampling data stored in the memory are erased, and the data is stored again at the next sampling time. Thereby, the total voltage Vt and the discharge current I in the latest battery state can be sampled, and the accurate maximum discharge power PD and maximum charge power PC can be calculated based on the sampling data in the latest battery state.

<< Charging / Discharging Power Calculation Process >> FIG. 15 is a flowchart showing the power calculation process of the battery controller 16. The battery controller 16 repeatedly executes this processing while the electric vehicle is in operation. In step 1, the total voltage Vt and the current I are sampled at a plurality of points when the discharge current increases as described above, and sampling data is stocked for each current region by the method described above. Next, in order to perform sampling for each predetermined amount of power,
In step 2, it is confirmed whether or not a predetermined amount of electric power has been discharged.
When the predetermined amount of electric power has been discharged, the process proceeds to step 3, otherwise the process returns to step 1. When the sampling is completed, it is confirmed in step 3 whether sampling data is stocked in three or more divided current regions in order to prevent regression calculation in a narrow current range and improve calculation accuracy.
If data is stocked in three or more areas, the process proceeds to step 4, and the VI characteristic is linearly regressed by the stock data. On the other hand, if the sampling data is not stocked in the three or more divided current regions in step 3, the maximum charging / discharging powers PD and PC are not calculated and the process returns to step 1.

In step 5, the current IDmax at the final discharge voltage Vmin is obtained from the regression line, and the above equation 5 is obtained.
The maximum discharge power PD is calculated by the following formula, the current ICmax at the allowable maximum voltage Vmax is calculated, and the maximum charge power PC is calculated by the above formula 4. In Step 6, Equation 9
Then, the cell variation correction coefficient M is obtained. In the subsequent step S7, if the correction coefficient M exceeds the predetermined value Mk, it is determined that an abnormality has occurred due to an increase in cell variation, and an alarm is issued in step 8. In step 9, the maximum charging power PC and the maximum discharging power PD are corrected by the cell variation correction coefficient M.
As is clear from Equations 4 and 5 above, the output power and the charging power of the battery 11 are inversely proportional to the internal resistance R, and therefore the maximum charging power PC and the maximum discharging power PD of the calculation result are obtained.
Is corrected as follows.

## EQU10 ## PC '= PC / M, PD '= PD / M

Output Limiting Processing Depending on the running pattern of the vehicle during power running, a running load exceeding the maximum discharge power PD '(= PD / M) may be applied. Maximum discharge power PD '
Is the power calculated based on the data sampled for a relatively short time from the rise of the discharge current, and can be said to be the power that can be discharged in a short time. For example, when a large current is discharged for a time longer than the sampling interval during power calculation, the maximum discharge power PD 'is exceeded and the total voltage Vt becomes equal to or lower than the discharge end voltage Vmin. Also, if the depth of discharge DOD increases, the maximum discharge power P
Since D'decreases, the same result is obtained when a large current is discharged for a time longer than the calculation interval of the maximum discharge power PD '. After a predetermined time from the rise of the discharge current,
If the V-I characteristic is sampled at a plurality of time points for a long time and the maximum discharge power PD 'is calculated based on the sampling result, the maximum discharge power PD' for a long time discharge can be calculated. However, since discharge for a long time is rare, the number of data after a long time from the rise of the discharge current is small, and if the regression line is calculated including such data, the accuracy of the regression line decreases. There is a risk of causing it to be undesirable. On the other hand, the calculated maximum discharge power PD 'may include some error. Maximum output is required according to the driving pattern of the vehicle, and when discharging to the maximum discharge power PD ',
If there is an error in the maximum discharge power PD ', the total voltage V
It is highly possible that t falls below the discharge end voltage Vmin.

Therefore, as shown in FIG. 16, the total voltage Vt
When the voltage becomes equal to or lower than the reference voltage V1, the maximum discharge power PD 'is corrected by the limiting coefficient K to limit the output. This output limitation is repeated every predetermined time T2 until the total voltage Vt becomes equal to or higher than the reference voltage V1. In FIG. 16, it is assumed that discharge is started at time t1 and the discharge power exceeds the maximum discharge power PD '. At time t2 when the total voltage Vt becomes equal to or lower than the reference voltage V1, the limiting coefficient K is updated from 1 to k. At time t3 after the control delay time T1, the maximum discharge power is k · PD ′
Limited to. As a result, the discharge current I decreases and the total voltage Vt increases. At time t4, which is T2 hours after time t2, the total voltage Vt is compared with the reference voltage V1, and V t <V1
If so, the limiting coefficient K is updated to limit the output, and V t ≧ V
If it is 1, the limiting coefficient K and the maximum discharge power PD 'are not changed. In this example, since V t <V1 at time t4,
The limiting coefficient K is k2. Time t after control delay time T1
At 5, the maximum discharge power is limited to k2 · PD ', the discharge current I decreases and the total voltage Vt increases. Next, at time t6, which is T2 hours after time t4, V t <V1 is satisfied, and therefore the limiting coefficient K is updated to k3. At time t7 after the control delay time T1, the maximum discharge power is limited to k3.PD '. As a result, the discharge current I decreases and the total voltage Vt increases. At time t8, which is T2 hours after time t6, the total voltage Vt is higher than the reference voltage V1.
Do not update.

During the period from time t1 to time t8 immediately after the start of discharge, the discharge power overshoots and the maximum discharge power P
D'is frequently restricted every T2 hours. As described above, the overshoot of the discharge power immediately after the start of discharge is due to the calculation error of the maximum discharge power PD '. On the other hand, at time t9 when the steady state is reached, V t <V1 is detected again, and the limiting coefficient K is updated to k4. At time t10 after the control delay time T1, the maximum discharge power is limited to k4 · PD ′, the discharge current I decreases, and the total voltage Vt increases. The excess of the discharge power in the steady state is because the discharge continued for a long time as described above. When the discharge mode is switched to the regenerative charge mode at time t12, the total voltage Vt rises sharply, and the limiting coefficient K is reset to 1 at this point.

The reference voltage V1 is

## EQU11 ## Any value that satisfies V1 ≧ Vmin can be selected. Further, the repeating time T2 of the output limiting process is set to be longer than the control delay time T1, and the constant k is set to an optimum value at which the overshoot of the discharge power becomes 0 within a predetermined convergence time. The minimum value of the limiting coefficient K is set to 0.1 and values below that are considered to be 0.
For example, the convergence time is 3S. When k = 0.8,
When n = 10, K becomes smaller than 0.1. Therefore, if the repetition time T2 is selected to be 300 mS, the limiting coefficient K becomes 0 and the output power becomes 0 at 3S.

FIG. 17 is a flowchart showing the output limiting process. The battery controller 16 executes this process when discharging is started. In step 11, if the total voltage Vt is smaller than the reference voltage V1 or the discharge stop signal is input from the cell controller 17, step 1
Then, the process proceeds to 2 and the output limiting process is performed. On the other hand, when the total voltage Vt is equal to or higher than the reference voltage V1 and the discharge stop signal is not input, the process proceeds to step 13. In step 13, it is confirmed whether the current I is 0 or negative, that is, whether the discharge mode is switched to the regenerative charge mode. If it remains in the discharge mode, the process returns to step 11, and if it switches to the regenerative charging mode, the process proceeds to step 18. In step 18, the variable n indicating the output limit number is reset to 0, the output limit coefficient K is reset to 1, and the process ends.

On the other hand, in step 12, the variable n indicating the output limit number is incremented. The initial value of the variable n is 0. In step 14, the limiting coefficient K is set. At the first output restriction, n = 1, so the restriction coefficient K
Is k. In step 15, the maximum discharge power PD 'in which the cell variation is corrected is multiplied by the limiting coefficient K for correction. In step 16, the timer is set to T2 and started. This T2 time is the repeating time of the output limiting process described in FIG. In step 17, it is confirmed whether the current I is 0 or negative, that is, whether the discharge mode is switched to the regenerative charge mode, and when the current is switched to the regenerative charge mode, the process proceeds to step 18, the variable n is reset to 0, and the limit is set. The coefficient K is reset to 1 and the process ends. On the other hand, when the discharge mode is continuing, the routine proceeds to step 19, where it is confirmed whether the timer has timed up and T2 time has elapsed. When the time T2 has elapsed, the process returns to step 11 and the above process is repeated.

<< Regeneration Restriction Process >> Similar to power running, regenerative power exceeding the maximum charging power PC '(= PC / M) may be generated during regenerative charging depending on the running pattern of the vehicle. The maximum charging power PC ′ is the power calculated based on the data sampled for a relatively short time from the rise of the discharge current, and can be said to be the power that can be regeneratively charged in a short time. For example, if regenerative charging with a large current is performed for a time longer than the sampling interval during power calculation, the maximum charging power PC will be exceeded and the total voltage Vt will exceed the maximum allowable voltage Vmax. Further, the calculated maximum charging power PC 'may include some errors. The maximum regenerative braking force is required according to the running pattern of the vehicle, and when charging is performed with the maximum charging power PC ', if the maximum charging power PC' has an error, the total voltage Vt exceeds the allowable maximum voltage Vmax. More likely.

Therefore, as shown in FIG. 18, the total voltage Vt
Exceeds the reference voltage V2, the maximum charging power PC 'is corrected by the limiting coefficient K to limit the regeneration. This regeneration limitation is repeated every predetermined time T2 until the total voltage Vt becomes equal to or lower than the reference voltage V2. In FIG. 18, time t1
It is assumed that the regenerative charging is started and the charging power exceeds the maximum charging power PC '. At time t2 when the total voltage Vt exceeds the reference voltage V2, the limiting coefficient K is updated from 1 to k. At time t3 after the control delay time T1, the maximum charging power PC ′ is k · PC ′.
Limited to. As a result, the charging current I and the total voltage Vt
Is reduced. At time t4 after time T2 from the time t2, to compare the total voltage Vt and the reference voltage V2, and updates the restriction coefficient K if V t> V2 limit output restriction coefficient if V t ≦ V2 K and maximum charging power PC 'are not changed. In this example, since V t > V2 at time t4, the limiting coefficient K is set to k2. Time t5 after control delay time T1
, The maximum charging power is limited to k2 · PC ', and the charging current I
And the total voltage Vt decreases. Next, from time t4 to T2
At time t6 after a lapse of time, since V t > V2, the limiting coefficient K is updated to k3. Time t7 after delay time T1
The maximum charging power is limited to k3 · PC ', and the discharge current I and the total voltage Vt decrease. At time t8, which is T2 hours after time t6, the total voltage Vt is lower than the reference voltage V2,
Therefore, the limiting coefficient K is not updated.

During the period from time t1 to time t8 immediately after the start of regenerative charging, the charging power overshoots and the maximum charging power PC 'is frequently limited every T2 hours. As described above, the overshoot of the charging power immediately after the start of the regenerative charging is due to the calculation error of the maximum charging power PC '. On the other hand, at time t9 when the steady state is reached, V t > V2 is detected again, and the limiting coefficient K is updated to k4. At time t10 after the control delay time T1, the maximum charging power is limited to k4 · PC ', and the charging current I and the total voltage Vt decrease. The excess of the charging power in the steady state is because the charging is continued for a long time as described above. When the regenerative charging mode is switched to the discharging mode at time t12, the total voltage Vt sharply drops, and the limiting coefficient K is reset to 1 at this point.

The reference voltage V2 is

(12) Any value that satisfies V2 ≦ Vmax can be selected. Further, the repetition time T2 of the regeneration limiting process is set to be longer than the control delay time T1, and the constant k is set to an optimum value at which the overshoot of the charging power becomes 0 within a predetermined convergence time. The minimum value of the limiting coefficient K is set to 0.1 and values below that are considered to be 0.
For example, the convergence time is 3S. When k = 0.8, K becomes smaller than 0.1 when n = 10. Therefore, if the repetition time T2 is selected to be 300 mS, the limiting coefficient K becomes 0 and the regenerative power becomes 0 at 3S. In the output limiting process and the regeneration limiting process described above, an example is shown in which the limiting process is performed at the same time interval T2, but the output limiting process and the regeneration limit may be performed at different time intervals. .

FIG. 19 is a flow chart showing the regeneration limiting process. The battery controller 16 executes this process when regenerative charging is started. In step 21,
If the total voltage Vt exceeds the reference voltage V2, or if an overvoltage signal is input from the cell controller 17, step 22
Proceed to and perform regeneration restriction processing. On the other hand, if the total voltage Vt is equal to or lower than the reference voltage V2 and no overvoltage signal is input, the process proceeds to step 23. In step 23, it is confirmed whether the current I is 0 or positive, that is, whether the regenerative charging mode is switched to the discharging mode. If it remains in the regenerative charging mode, the process returns to step 21, and if it switches to the discharge mode, the process proceeds to step 28. In step 28, the variable n indicating the number of times of regeneration limitation is reset to 0, the limiting coefficient K is reset to 1, and the process ends.

On the other hand, in step 22, the variable n indicating the number of times of regeneration limitation is incremented. The initial value of the variable n is 0. In step 24, the limiting coefficient K is set. At the first output restriction, n = 1, so the restriction coefficient K
Is k. In step 25, the maximum charging power PC 'in which the cell variation has been corrected is multiplied by the limiting coefficient K to make the correction. In step 26, the timer is set to T2 and started. This T2 time is the repetition time of the regeneration limiting process described in FIG. In step 27, it is confirmed whether the current I is 0 or positive, that is, whether the regenerative charging mode is switched to the discharging mode. When the discharging mode is switched to, the process proceeds to step 28, the variable n is reset to 0, and the limiting coefficient is set. The K is reset to 1 and the process ends. On the other hand, when the regenerative charging mode continues, the routine proceeds to step 29, where it is confirmed whether the timer has timed up and T2 time has elapsed. When T2 time has elapsed, the process returns to step 21 and the above process is repeated.

As shown in FIG. 2, the DOD is 60%.
If it exceeds, the internal resistance R during discharging and the internal resistance R during charging
, The calculated maximum charging power PC includes a large error. However, when the DOD exceeds 60%, since the true maximum charging power of the battery is sufficiently larger than the maximum power regenerated from the inverter 12, even if there is a large error in the calculated value PC of the maximum charging power, there is a problem. I won't.

In the configuration of the above embodiment, the voltage sensor 14 is the voltage measuring means, the current sensor 15 is the current measuring means, and the battery controller 16 and the cell controller 17 are the power calculating means , the power correcting means and the alarm.
Each means is configured.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an electric vehicle according to an embodiment.

FIG. 2 is a diagram showing the relationship between the depth of discharge (DOD) of a battery and the internal resistance.

FIG. 3 is a diagram showing changes in the total voltage of the battery when discharged with a constant current.

FIG. 4 is a diagram showing a regression line based on sampling data of total voltage and discharge current during discharging.

FIG. 5 is a diagram showing the relationship between cell voltage and discharge current.

FIG. 6 is a diagram illustrating sampling timing of total voltage and discharge current during discharging.

FIG. 7 is a diagram showing a case where the discharge current and the total voltage are sampled after a predetermined time has elapsed from the rise of the discharge current.

FIG. 8 is a diagram in which the sampling data shown in FIG. 7 is plotted on a VI graph and linear regression is performed.

FIG. 9 is a diagram showing a case where the discharge current and the total voltage are sampled after a predetermined time Δt1 and Δt2 from the rising of the discharge current.

10 is a diagram in which the sampling data shown in FIG. 9 is plotted on a VI graph and linear regression is performed.

FIG. 11 is a diagram showing an example in which VI characteristics are sampled at a plurality of points in a normal traveling pattern of an electric vehicle.

FIG. 12 shows the sampling data shown in FIG.
It is the figure which plotted on the graph and performed linear regression.

FIG. 13 is a diagram in which the sampling data shown in FIG. 11 is classified according to sampling timing and discharge current.

FIG. 14 is a diagram for explaining a stocking method of sampling data.

FIG. 15 is a flowchart showing a power calculation process.

FIG. 16 is a diagram for explaining output restriction.

FIG. 17 is a flowchart showing an output limiting process.

FIG. 18 is a diagram for explaining regenerative restriction.

FIG. 19 is a flowchart showing a regeneration limiting process.

[Explanation of symbols]

11 batteries 12 inverter 13 motor 14 Voltage sensor 15 Current sensor 16, 16A, 16B controller 17 cell controller 111-11n single cell

─────────────────────────────────────────────────── --- Continuation of the front page (56) Reference JP-A-9-312901 (JP, A) JP-A-9-171063 (JP, A) JP-A-9-84205 (JP, A) JP-A-9- 218251 (JP, A) JP-A-6-34727 (JP, A) JP-A-56-126777 (JP, A) JP-A-9-215111 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) H02J ^7/ 00-7/04 B60L 3/00 B60L 11/18

Claims (3)

(57) [Claims] 1. A voltage measuring means for measuring a terminal voltage of an assembled battery in which a plurality of unit cells are connected in series, a current measuring means for measuring a current flowing through the assembled battery, the voltage measuring means and the current measurement. Measured by means
Linear regression of multiple measurement results on a two-dimensional plane of voltage and current
Then, the voltage axis intercept of this regression line is the open circuit voltage of the assembled battery.
Eo, the slope of the regression line and the internal resistance R of the battery pack
Of the open circuit voltage Eo, the internal resistance R, and the assembled battery
Maximum allowable voltage Vmax and discharge end voltage Vm of the assembled battery
Maximum charging power PC and maximum discharging of the assembled battery based on in
A power calculation means for calculating the power PD, and an average value Rave and a maximum value Rm of the internal resistances of the plurality of unit cells.
The ratio Rmax / Rave with ax is set as the correction coefficient M, and the maximum charge
Power PC to PC / M, maximum discharge power PD to PD / M
A power control device for an electric vehicle, comprising: power correction means for correcting each . 2. The electric power control device for an electric vehicle according to claim 1, wherein the electric power correction means sets the open circuit voltage Eo and the assembled battery
Terminal voltage, specified voltage of the unit cell and number of units of the unit cell
An electric power control device for an electric vehicle, wherein the correction coefficient M is calculated based on a number . 3. The electric power control device for an electric vehicle according to claim 1, wherein an alarm hand is provided to give an alarm when the correction coefficient M exceeds a predetermined value.
A power control device for an electric vehicle, characterized in that it comprises a step .