JP5191014B2 - Battery state quantity calculation method - Google Patents

Description

  The present invention relates to a battery state quantity calculation method, and more particularly to a battery state quantity calculation method for calculating a battery state quantity using at least a plurality of voltage / current pairs detected from the battery.

  A battery used for a vehicle needs to estimate the state of the battery with high accuracy from the necessity of capacity management and safety management. Conventionally, for this purpose, various battery state quantity computing devices for computing battery state quantities (for example, battery voltage, current, remaining capacity, open circuit voltage, internal resistance, etc.) have been proposed or put into practical use.

  The configuration common to these battery state quantity calculation devices collects a large number of battery voltage / current pairs as input parameters, and uses the relationship between the voltage and current obtained based on them to estimate the battery state quantity. That is.

  However, since there are many phenomena related to the relationship between voltage and current, such as the remaining capacity state, temperature, deterioration state, and polarization state of the battery, the relationship between voltage and current is a simple linear relationship did not become.

  One method of estimating the relationship between voltage and current used in the past is to create a regression line by plotting multiple voltage / current pairs in voltage (V) and current (I) in a two-dimensional space. A method of using this regression line to estimate a current value corresponding to an input voltage value or a voltage value corresponding to an input current value (regression line utilization method) is widely used at present. The slope of the regression line is a value indicating the internal resistance of the battery.

  In addition, a large number of sampled voltage / current pairs and other battery state quantities (such as temperature and operating time) may be input to the neural network to estimate battery state quantities (such as SOC) as output parameters. It has been proposed (Patent Documents 1 and 2).

JP-A-9-243716 JP 2003-249271 A

  However, in the estimation of the battery state quantity using the above-described conventional regression line utilization method, since the coordinate points of a large number of voltage / current pairs are widely dispersed in the two-dimensional space, the accuracy of the obtained regression line is low. It was widely known that the battery state quantity error obtained using the line was very large.

  In addition, the above-described neural network calculation method can obtain satisfactory calculation accuracy even though the processing device and the processing load are significantly increased as compared with the battery state quantity estimation using the conventional regression line utilization method. The fact was that it was not possible.

  Next, considering a vehicle power supply system, a large fluctuation in power supply voltage causes various problems such as unstable load operation, flickering of headlamps, and a decrease in battery life. Therefore, there is a strong demand to stabilize the power supply voltage (battery voltage) in spite of changes in load conditions. In conventional power supply voltage management in a vehicle, feedback control is generally performed so that the difference between a detected power supply voltage and its target voltage converges to zero. However, if the power supply voltage is changed by this feedback control, the internal state of the battery changes accordingly, so that hunting occurs or delays in the power supply voltage, so that it can respond quickly and quickly to sudden changes in the power supply voltage due to fluctuations in the electrical load, for example. Therefore, it is difficult to perform stabilization control.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide a battery state quantity calculation method capable of improving the calculation accuracy of the battery state quantity of the battery.

  A battery state quantity calculation method according to a first aspect of the present invention that solves the above problem is a regression line that defines a relationship between a voltage and a current of a battery from a plurality of voltage / current pairs detected from a battery connected to a circuit system. In the battery state calculation method of calculating a battery state quantity that is an electric quantity indicating the state of the battery based on the regression line, the regression is performed so as to pass through the coordinate point of the voltage / current pair collected immediately before. A line is shifted, and a battery state quantity of the battery is calculated based on the shifted regression line.

  That is, according to the present invention, when one of the voltage and the current is input to the regression line indicating the relationship between the voltage and the current and the other is calculated, the coordinates of the voltage / current pair collected immediately before the regression line are calculated. Since the shift is made so that it passes through a point, and either the voltage or current as an input parameter is input to this shifted regression line and the other is extracted, the previous battery state can be reflected well on the regression line. In addition, the calculation error can be greatly reduced and the battery state quantity can be calculated with high accuracy. This shift means that when the regression line is a straight line, the regression line is drawn from the coordinate position of the previous voltage / current pair at the inclination angle of the regression line.

  In a preferred aspect, the shift is performed so that the shift distance is minimized. That is, the shift may be a movement in the direction parallel to either the voltage axis or the current axis (axis translation), but instead, on the regression line closest to the coordinate position of the previous voltage / current pair. Since the amount of movement of the regression line can be reduced the most by shifting the coordinate point to the coordinate position of the immediately preceding voltage / current pair, the calculation error can be reduced.

  In a preferred aspect, the battery current is determined from the coordinates of the intersection of the shifted regression line and a straight line indicating a predetermined voltage value of the battery. In this way, it is possible to satisfactorily obtain the value of the battery charge / discharge current required when the battery voltage is set to the predetermined target voltage, and to control the battery charge / discharge current as described above to obtain the vehicle power supply voltage. It can be satisfactorily applied when controlling.

  In a preferred aspect, the voltage of the battery is determined from the coordinates of the intersection of the shifted regression line and a straight line indicating a predetermined current value of the battery. If it does in this way, the battery voltage with respect to the value of the predetermined charging / discharging electric current of a battery can be calculated | required accurately. For example, if the charge / discharge current value is 0 as the predetermined charge / discharge current value, the open circuit voltage value of the battery often used in the battery state calculation can be obtained with high accuracy.

  A battery state quantity calculation method according to a second aspect of the present invention for solving the above problem is a regression line for defining a relationship between a voltage and a current of a battery from a plurality of voltage / current pairs detected from a battery connected to a circuit system. The battery state quantity obtained from the regression line using a plurality of regression lines in a battery state computation method for computing a battery state quantity that is an electrical quantity indicating the battery state based on the regression line And calculating the calculation error which is the difference between the measured battery state quantity and the regression line regarded as having a small calculation error among the plurality of regression lines, for the next battery state quantity. It is said.

  That is, according to the present invention, control is performed using the voltage or current value, which is the output parameter obtained using the regression line, to detect the actual value of the output parameter, and the actual value and the regression line are used. By using the calculation error between the calculated values, the regression line that is considered to have a small calculation error is selected from the multiple regression lines for the next calculation using the regression line. The error can be greatly reduced.

  In other words, although the regression line depends on the voltage / current pair group sampled in the past, the internal state of the battery at each sampling time is different, so the voltage and current between the voltage and current corresponding to the current internal state of the battery are accurate. It does not represent a relationship. Therefore, in a preferred aspect, when the calculation error when one of the plurality of regression lines is used exceeds a predetermined threshold value, another one of the plurality of regression lines is used as the next battery. Select for state quantity. In this way, the regression line can be easily switched.

  A battery state quantity calculation method according to a third aspect of the present invention for solving the above-described problem is a regression line for defining a relationship between a voltage and a current of a battery from a plurality of voltage / current pairs detected from a battery connected to a circuit system. And calculating the battery state quantity obtained from the regression line and the actually measured battery state quantity in a battery state calculation method for calculating a battery state quantity that is an electric quantity indicating the state of the battery based on the regression line And calculating the regression line for the next calculation to a direction in which the calculation error is considered to be small based on the calculation error.

  That is, according to the present invention, control is performed using the voltage or current value, which is the output parameter obtained using the regression line, to detect the actual value of the output parameter, and the actual value and the regression line are used. Using the calculation error between the calculated values, the characteristics of the regression line are corrected in the direction that the calculation error is considered to be smaller before the next calculation using the regression line. Can be greatly reduced.

  In a preferred aspect, the slope of the regression line is changed in a direction that reduces the calculation error. In this way, correction can be easily performed.

  A battery state quantity calculation method according to a fourth aspect of the present invention for solving the above-described problem is a regression line that defines the relationship between the voltage and current of the battery from a plurality of voltage / current pairs detected from the battery connected to the circuit system. And calculating a battery state quantity, which is an electric quantity indicating the state of the battery based on the regression line, in a battery state calculation method, wherein the regression line is a regression line when the battery is in a discharged state. And a regression line at the time of charging that is the regression line when the battery is in a charged state are determined separately, and when the current state of the battery is in a discharged state, based on the regression line at the time of discharge The battery state quantity of the battery is calculated, and when the current state of the battery is a charged state, the battery state quantity of the battery is calculated based on the regression line at the time of charging. There. In this way, since the regression lines are used separately at the time of charging and discharging, where the characteristics of the regression lines are significantly different due to reasons such as different polarization states, the calculation error of the battery state quantity can be reduced.

  In a preferred embodiment, a discharge current increase regression line which is the regression line when the battery is in a discharge state and a current increase tendency, and a discharge current decrease regression which is the regression line when the battery is in a discharge state and a current decrease tendency. A charging current increase regression line that is the regression line when the battery is in a charged state and a current increasing tendency; and a charging current decrease regression line that is the regression line when the battery is in a charged state and a current decreasing tendency; Are determined separately, and when the current state of the battery is an increased discharge current state, the battery state amount of the battery is calculated based on the regression line when the discharge current increases, and the current state of the battery is discharged. When the battery is in a current decreasing state, the battery state amount of the battery is calculated based on the regression line at the time of discharging current decreasing, and when the current state of the battery is a charging current increasing state The battery state amount of the battery is calculated based on the charging current increase regression line, and when the current state of the battery is a charging current decrease state, the battery state of the battery based on the charging current decrease regression line is calculated. Calculate the quantity.

  In this way, since the regression line is used separately when the current increases and decreases when the voltage / current characteristics of the battery differ due to the influence of polarization or the like, the calculation error of the battery state quantity can be reduced. .

  In addition, the calculation of the battery state quantity based on the regression line in each of the above-described inventions includes using a characteristic function that defines the regression line (for example, the internal resistance value of the battery) for the calculation of the battery state quantity.

1 is a block circuit diagram illustrating a vehicle power supply voltage control apparatus according to a first embodiment. 3 is a flowchart illustrating vehicle power supply voltage control according to the first embodiment. It is a block circuit diagram which shows the modification of FIG. 2 is a flowchart illustrating a control current Is calculation method of FIG. 1. 2 is a flowchart illustrating a control current Is calculation method of FIG. 1. 2 is a flowchart illustrating a control current Is calculation method of FIG. 1. It is explanatory drawing which shows the example of the shift of a regression line, and an angle change. 6 is a flowchart illustrating a control current Is calculation method according to the second embodiment. It is a flowchart which shows the regression line update method of Embodiment 2.

  Hereinafter, preferred embodiments of the present invention will be described. However, the present invention is not limited to the following embodiments, and the technical idea of the present invention may be implemented using other known techniques.

(Embodiment 1)
A vehicle power supply voltage control apparatus using the battery state quantity calculation method of this embodiment will be described with reference to FIG.

(overall structure)
1 is a battery, 2 is a bidirectional current control device, 3 is a battery controller, and 4 is a vehicle ECU. The battery 1 is connected to the power supply line 5 through the bidirectional current control device 2, and is connected to each on-vehicle electric load or generator (not shown) through the power supply line 5. Reference numeral 6 denotes a current sensor that detects a charge / discharge current of the battery 1.

  The voltage V and current I of the battery 1 and the power supply voltage VL which is the voltage of the power supply line 5 are input to the battery controller 3 built in the microcomputer. The battery controller 3 controls the power supply voltage VL by controlling the charge / discharge current of the battery 1 through the bidirectional current control device 2 based on the target value (also referred to as control voltage) Va of the power supply voltage VL specified by the vehicle ECU 4. Power supply voltage stabilization control is performed to converge to the voltage Va. In the following description, since the charge / discharge current of the battery 1 is controlled by the bidirectional current control device 2, it is also referred to as a control current Is.

(Power supply voltage stabilization control)
The power supply voltage stabilization control operation performed by the battery controller 3 will be described with reference to FIG.

  First, the voltage V, current I, control voltage Va of the battery 1 and the power supply voltage VL which is the voltage of the power supply line 5 are read (S1000) and stored in advance or between the voltage and current of the battery 1 created this time. The control voltage Va, which is the target value of the power supply voltage, is substituted into the relationship, and the control current Is, which is the target value of the current of the battery 1 corresponding to the control voltage Va, is calculated (S2000). It transmits to the current control device 2 (S3000). The bidirectional current control device 2 is, for example, a bidirectional current control circuit, and performs switching control of the built-in switching element so that the charge / discharge current of the battery 1 becomes the control current Is. Thereafter, it is determined whether or not the vehicle has finished traveling (S4000). If the vehicle is traveling, the process returns to step S1000. The battery controller 3 has an interrupt routine that is performed periodically or in a predetermined battery state in order to obtain the above-described calculation function of the control current Is in advance. Based on this, processing for selecting a calculation function of the control current Is with a small calculation error is performed. Accordingly, step S2000 is a process of reading the value of the control current Is calculated in the interrupt routine.

(Modification)
FIG. 3 shows a modification of the vehicle power supply voltage control apparatus shown in FIG. In this modification, the charging / discharging current control of the battery 1 that is performed by the bidirectional current control device 2 in FIG. 1 is performed on the generator 7 side, thereby omitting the bidirectional current control device 2. is there.

  The power supply voltage stabilization control operation of this modification will be described. First, the voltage V of the battery 1 (= power supply voltage VL), the current I, and the output current (power generation current) IG of the generator 7 are read, and then stored in advance or the voltage and current of the battery 1 created this time Substituting the control voltage Va, which is the target value of the power supply voltage, into the relationship between them, the control current Is, which is the target value of the current of the battery 1, is calculated, and then the control current Is, the generated current IG, and the measured current I of the battery 1 are calculated. Then, the generated current IGS to be output next time by the generator 7 is calculated from the following equation.

IGS = (IG-I) + Is
That is, (IG-I) is the total current consumption of the in-vehicle electric load, and the generator 7 is instructed to output the generated current IGS so that the next generated current IGS matches the total current control current Is. To do. The regulator built in the generator 7 calculates a field current value corresponding to the commanded generation current IGS, and a field current corresponding to the field current value is supplied to the field coil. Thereby, the vehicle power supply voltage control which considered the characteristic of the battery 1 similarly to the said Embodiment 1 is realizable.

(Description of control current Is determination method using regression line)
(Regression line selection)
Processing for calculating the control current Is used in step S2000 will be described with reference to FIG.

  In this embodiment, a plurality of calculation functions (control current Is calculation formula), that is, regression lines defined by regression formulas are used, and a suitable one of them is selected according to the situation, and the control current Is is selected. Perform the calculation. Further, in addition to simply adding the current value of the voltage / current pair in calculating a new calculation function, the past calculation function is also corrected according to the battery state. Each control current Is calculation formula used in the following examples uses a regression line, but is not limited thereto. For example, a regression curve having a predetermined curvature may be created on a two-dimensional plane on which a large number of voltage / current pairs are plotted, and the control current Is may be calculated by substituting the target voltage into this regression curve. The current value at the intersection coordinate between the tangent line of the upper measured coordinate point and the line indicating the target voltage may be used as the control current Is.

  In FIG. 4, the routine is started by starter start (S100), and a large number of voltage / current pairs are measured during a large current change period of the battery 1 by starter start (S101). Next, the internal resistance R of the battery 1 is calculated and stored using a number of voltage / current pairs collected during the engine starting period (S102).

  The calculation process of the internal resistance R in step S102 will be described in more detail with reference to FIG. By averaging a group of voltage / current pairs using a number of voltage / current pair groups acquired at starter start, noise components at the time of measurement are removed (S201). Next, the current and voltage range can be calculated so that the battery internal resistance can be calculated so that the difference between the estimated voltage value and the measured voltage value can be minimized when predicting the voltage drop at start-up. A voltage / current pair for recovery after the voltage reaches the lower limit is selected from the voltage / current pair at the time of first flowing into the starter (S202). Next, a regression line (starting regression line) is created from the selected voltage / current pair group by a known method, and the slope is calculated and stored as an internal resistance (S204).

  Thereafter, the voltage / current pair is collected again at the start of traveling (S103), and it is checked whether or not the flag for selecting the control current Is calculation method (referred to as the internal resistance use control current calculation method) is 0 (S104). Here, the flag value 0 indicates that the internal resistance utilization control current calculation method shown in steps S105 to S107 described later should be used for the calculation of the control current Is, and the flag value 1 indicates steps S111 to S11 described later. It shows that the regression line utilization control current calculation method shown in S113 should be used for calculation of the control current Is. Immediately after traveling, the flag is reset to 0 by a reset initialization step performed immediately after the start of the routine. Therefore, immediately after traveling, the process proceeds to step S105.

  In step S105, a regression line (running regression line) is created from the voltage / current pair collected during traveling in step S103, as in step S102, and the internal resistance R during traveling is determined from the slope of the traveling regression line. Is calculated and stored. Note that when the voltage / current pair having sufficient dispersion to generate a regression line with high accuracy with little change in the voltage and current of the battery 1 during traveling cannot be adopted, the startup regression line may be used.

  Next, the control voltage Va which is the target value of the power supply voltage VL of the power supply line 5 read from the vehicle ECU 4 is read (S106), and the control current Is, that is, the control current Is calculated from the read control voltage Va and the internal resistance R calculated in step S105. The target value of the charging / discharging current of the battery 1 is calculated, the control current Is is commanded to the bidirectional current control device 2, and the charging / discharging current of the battery 1 is controlled to the control current Is (S107). The calculation of the control current Is is performed by the following equation, for example. V is the battery voltage data collected immediately before.

Is = (V−Va) / R
The bidirectional current control device 2 that controls the charging / discharging current of the battery 1 has an internal resistance, but the internal resistance R of the bidirectional current control device 2 is relatively small and is ignored. That is, since the bidirectional current control device 2 is opened and closed with a predetermined duty at a predetermined cycle like a switching regulator, for example, the on-resistance loss when closed, that is, when the current is energized is small and can be ignored.

  Next, a voltage / current pair is sampled again, a voltage error α between the sampled voltage V and the control voltage Va is calculated (S108), and it is determined whether the voltage error α is equal to or less than a predetermined threshold value (S109). . If the voltage error α is less than the threshold value, it is determined that the current control current Is calculation method is small and should be used in the future, and the control current Is calculation method using the travel regression line shown in steps S103 to S107 (internal) 0 is selected as a flag value for adopting the resistance utilization control current calculation method (S110), and it is determined whether or not the running is finished (S117). If the running is finished, the routine is finished and finished. If not, the process returns to step S103. Next, if the flag is 1 in step S104, the process proceeds to step S111 in order to use the second regression line utilization control current calculation method shown in steps S111 to S113 for calculation of the control current Is.

  In step S111, a voltage / current pair group that is different from the voltage / current pair group selected for the generation of the regression line at the time of traveling in step S105 is selected from the voltage / current pair group collected in steps S101 and S103. Create 2 regression lines. As the voltage / current pair group to be selected for the second regression line, for example, only the immediately preceding voltage / current pair group may be selected, or all voltage / current pairs may be adopted. Furthermore, a voltage / current pair group collected in an operation mode approximate to the current operation mode of the battery 1 may be employed. The operation mode that approximates the current operation mode of the battery 1 is divided into a charging current increasing state, a charging current decreasing state, a discharging current increasing state, a discharging current decreasing state, and the like. A voltage / current pair group that belongs to the same state may be employed, or a voltage / current pair group that approximates a separately calculated remaining capacity value may be employed.

  Next, the control voltage Va which is the target value of the power supply voltage VL of the power supply line 5 read from the vehicle ECU 4 is read (S112), and the read control voltage Va is substituted into the second regression line to obtain the control current Is, that is, the battery. 1 is calculated, and the control current Is is commanded to the bidirectional current control device 2 to control the charge / discharge current of the battery 1 to the control current Is (S113).

  Next, a voltage / current pair is sampled again, a voltage error α between the sampled voltage V and the control voltage Va is calculated (S114), and it is determined whether the voltage error α is equal to or less than a predetermined threshold value (S115). If the voltage error α is less than the threshold value, it is determined that the current control current Is calculation method has a small error and should be adopted in the future, and the control current Is calculation method shown in steps S103 to S107 (using the second regression line) In order to adopt the control current Is calculation formula), the flag value is set to 1 (S116). If it is determined whether or not the traveling is finished, the routine is finished, and if not finished, the process proceeds to step S103.

  If the voltage error α is greater than or equal to the threshold value in step S115, the process proceeds to step S118 shown in FIG.

(Revision line correction)
In step S118, the second regression line calculated in step S111 is translated (shifted) to the coordinate point (Vx, Ix) of the immediately preceding voltage / current pair to obtain a third regression line (immediately shifted regression line). To do. Next, the control voltage Va is read (S119), and the control current Is is calculated by substituting the control voltage Va into the previous shift regression line. If the inclination angle of the second regression line, that is, the internal resistance is R, the calculation in step S120 is equivalent to the calculation of the following equation.

Is = (Vx−Va) / R + Ix
Next, a voltage / current pair is sampled again, a voltage error α between the sampled voltage V and the control voltage Va is calculated (S121), and it is determined whether the voltage error α is equal to or less than a predetermined threshold value (S122). . If the voltage error α is less than the threshold, it is determined that the current control current Is calculation method has a small error and should be used in the future, and the process proceeds to step S116. If the voltage error α is greater than or equal to the threshold, the process proceeds to step S123.

  In step S123, the inclination angle of the previous shift regression line, that is, the internal resistance is changed by a small amount so that the voltage error α decreases. This change will be specifically described with reference to an explanatory diagram shown in FIG.

  A third regression line (immediate shift regression line) is obtained by parallel translation (shift) to the coordinate point (Vx, Ix) of the immediately preceding voltage / current pair. In FIG. 7, the control current Is obtained by substituting the control voltage Va for the previous shift regression line, and the control in which the measured voltage Vn of the battery 1 obtained after controlling the current of the battery 1 to the control current Is is the target voltage. When the voltage Va is smaller than the voltage Va (see FIG. 7), the inclination angle of the immediately preceding shift regression line m is set so that the control current Is increases, that is, the inclination angle of the immediately preceding shift regression line m decreases. Decrease and create a new regression line m + 1 and store it. This new regression line m + 1 is used in step S112 in place of creating the second regression line in the next step S111. Of course, when the actually measured voltage Vn of the battery 1 is larger than the control voltage Va, which is the target voltage, this previous shift regression is performed in the direction in which the control current Is decreases, that is, in the direction in which the inclination angle of the previous shift regression line m decreases. A new regression line m + 1 is created and stored by decreasing the inclination angle of the line m.

  However, the correction of the second regression line shown in FIG. 6 is sometimes stopped, and a new regression line equation is sometimes created using the new voltage / current pair obtained after the creation of the previous regression line in step S112. It is preferable to create it.

(Modification)
In the above-described embodiment, a regression line is used as a regression line, but a nonlinear regression curve may be adopted. In order to shift this regression curve to the coordinate point of the previous voltage / current pair, it is better to translate the voltage curve and current axis twice in parallel so that the shift distance is minimized on the two-dimensional plane. Yes.

(effect)
In this embodiment described above, the power supply voltage fluctuation of the vehicle power supply system, which has a large voltage change depending on the driving situation, is performed by controlling the charge / discharge current of the battery to the predetermined target value. Since the power supply voltage can be adjusted in consideration of the state change, the control response can be improved.

  Moreover, in the said embodiment, since the regression line which shows the relationship between a voltage and an electric current is shifted so that it may pass through the coordinate point of the voltage / current pair collected immediately before, a calculation error can be reduced.

  In the above embodiment, the output parameter obtained using the created regression line is created with a group of voltage / current pairs different from the original regression line when the calculation error between the measured value and the measured value is large. Since computation is performed using another regression line obtained by selecting a voltage / current pair group or a regression line created by a method different from the original regression line, computation errors can be reduced.

  Further, according to the above embodiment, when the calculation error is large, the regression line for the next calculation, in particular, the inclination angle (internal resistance) is corrected to a direction in which the calculation error is considered to be small. Can be reduced.

(Embodiment 2)
Another embodiment will be described. In this embodiment, a regression line is determined, and a regression line when the battery is in a discharged state and a current increasing tendency is a regression line when the battery is in a discharged state and a current is decreasing. The regression line when the discharge current decreases, the regression line when the battery is charged and the current increases, and the regression line when the battery is charged and the current decreases. The regression line when the battery is charged and the current decreases. And determine which of the above four states (referred to as operation mode) corresponds to the current battery state, and use a regression line that matches the current operation mode. It is characterized by calculating the battery state.

  This will be described more specifically with reference to the flowchart shown in FIG.

  First, a voltage / current pair is measured (S300), and the current battery operation mode is determined using a plurality of voltage / current pairs immediately before including the measured voltage / current pair (S302). In this embodiment, four operation modes of a discharge current increase state, a discharge current decrease state, a charge current increase state, and a charge current decrease state are set as the operation modes. Next, a regression line that matches the operation mode determined in step S302 is selected from the four regression lines stored in advance (S304). The four regression lines include a discharge current increase regression line, a discharge current decrease regression line, a charge current increase regression line, and a charge current decrease regression line. Next, the regression line selected this time is shifted to the coordinate position indicated by the voltage / current pair measured this time by the method described above (S306). More specifically, a straight line having a slope of the selected regression line passing through the coordinate position indicated by the voltage / current pair measured this time is drawn on the voltage / current plane, and the target voltage value is met on this straight line. The current value at the coordinate position to be determined is determined as the control current Is and output.

  In this way, since the regression line used is a regression line created in the same operation mode as the current battery operation mode, the calculation error can be greatly reduced.

  Next, a method for creating the four regression lines will be described with reference to the flowchart shown in FIG.

  First, the current operation mode is determined from a plurality of voltage / current pairs collected during the immediately preceding predetermined period (S400). Next, the four regression lines that have already been stored as the regression line that matches the determined operation mode (return line when discharge current increases, regression line when discharge current decreases, regression line when charge current increases, and regression line when charge current decreases) The selected regression line is updated (S402).

  The update of the regression line will be specifically described below by taking the regression line when the discharge current increases as an example. The voltage / current pairs collected in the past are grouped into the above four operation modes. A characteristic curve in this operation mode (when the discharge current is increased) is newly created by mixing multiple voltage / current pairs collected during the previous specified period with the voltage / current pair in the same operation mode (in this case, the discharge current is increased). To decide. When mixing the voltage / current pairs collected this time, for example, the data of the oldest voltage / current pair can be excluded. In this way, the regression line to be stored can be updated. Other operation modes can be updated in the same manner. As a specific calculation method for determining the regression line, in addition to the above, there is a calculation method for determining the next regression line more easily from the previous regression line and the voltage / current pair added this time. Since it is only the above problem, description is abbreviate | omitted.

DESCRIPTION OF SYMBOLS 1 Battery 2 Bidirectional current control device 3 Battery controller 5 Power supply line 7 Generator

Claims (10)

An electrical line indicating a state of the battery based on the regression line is determined by determining a regression line that defines a relationship between the voltage and current of the battery from a plurality of voltage / current pairs detected from the battery connected to the circuit system. In a battery state calculation method for calculating a battery state quantity that is a quantity,
Shift the regression line so that it passes through the coordinate point of the voltage / current pair collected immediately before,
A battery state calculation method, wherein a battery state quantity of the battery is calculated based on the shifted regression line. The battery state calculation method according to claim 1,
A battery state calculation method for performing the shift so that the shift distance is minimized. The battery state calculation method according to claim 1,
A battery state calculation method for obtaining a current of the battery from coordinates of an intersection of the shifted regression line and a straight line indicating a predetermined voltage value of the battery. The battery state calculation method according to claim 1,
A battery state calculation method for obtaining a voltage of the battery from coordinates of an intersection of the shifted regression line and a straight line indicating a predetermined current value of the battery. An electrical line indicating a state of the battery based on the regression line is determined by determining a regression line that defines a relationship between the voltage and current of the battery from a plurality of voltage / current pairs detected from the battery connected to the circuit system. In a battery state calculation method for calculating a battery state quantity that is a quantity,
Using a plurality of regression lines,
Calculating a calculation error that is a difference between the battery state quantity obtained from the regression line and the actually measured battery state quantity;
A battery state calculation method, wherein the regression line regarded as having a small calculation error among the plurality of regression lines is selected for the next battery state quantity. The battery state quantity calculation method according to claim 5,
When the calculation error when using one of the plurality of regression lines exceeds a predetermined threshold, the other one of the plurality of regression lines is selected for the next battery state quantity. A battery state calculation method. An electrical line indicating a state of the battery based on the regression line is determined by determining a regression line that defines a relationship between the voltage and current of the battery from a plurality of voltage / current pairs detected from the battery connected to the circuit system. In a battery state calculation method for calculating a battery state quantity that is a quantity,
Calculating a calculation error that is a difference between the battery state quantity obtained from the regression line and the actually measured battery state quantity;
A battery state calculation method, wherein the regression line for the next calculation is corrected to a direction in which the calculation error is considered to be small based on the calculation error. The battery state calculation method according to claim 7,
A battery state calculation method of changing the slope of the regression line in a direction in which the calculation error is reduced. An electrical line indicating a state of the battery based on the regression line is determined by determining a regression line that defines a relationship between the voltage and current of the battery from a plurality of voltage / current pairs detected from the battery connected to the circuit system. In a battery state calculation method for calculating a battery state quantity that is a quantity,
The discharge regression line that is the regression line when the battery is in the discharged state and the charge regression line that is the regression line when the battery is in the charged state are separately determined,
When the current state of the battery is a discharged state, the battery state amount of the battery is calculated based on the regression line at the time of discharge, and when the current state of the battery is a charged state, A battery state calculation method comprising: calculating a battery state quantity of the battery based on the battery state amount. The battery state calculation method according to claim 9,
A regression line at the time of discharge current increase, which is the regression line when the battery is in a discharged state and a current increase tendency;
A regression line at the time of discharge current reduction, which is the regression line when the battery is in a discharged state and current tends to decrease;
A charging current increase regression line that is the regression line when the battery is in a charged state and current increases, and
Regression line at the time of charge current decrease that is the regression line when the battery is in a charged state and current decreases, and
Are determined separately,
When the current state of the battery is a discharge current increasing state, the battery state amount of the battery is calculated based on the regression line when the discharge current increases, and when the current state of the battery is a discharge current decreasing state The battery state amount of the battery is calculated based on the regression line when the discharge current decreases, and the battery state of the battery based on the regression line when the charge current increases when the current state of the battery is a charge current increase state A battery state calculation method for calculating a battery state amount of the battery based on the charge current decrease regression line when the current state of the battery is a charge current decrease state.

Images