JP6221851B2 - Voltage converter - Google Patents

Description

  The present invention relates to a voltage converter that converts an input voltage from a battery into a system voltage that is output to a load.

2. Description of the Related Art Conventionally, there is known a voltage conversion device that converts a DC voltage of a battery by switching a current flowing through a reactor and outputs the converted voltage to a load such as an inverter that drives an AC motor.
In general, in this type of voltage converter, a duty command value, which is an on-time ratio with respect to a switching cycle, is calculated by feedback control so that a system voltage output to a load matches a command voltage (target voltage).

However, when feedback control of the voltage conversion device is performed by PI control, if the same control gain is used for the entire range of operating voltage, the controllability differs depending on the state of the voltage conversion device and the load, so that the voltage conversion device is output to the load. System voltage may fluctuate.
Therefore, for example, in the voltage converter of Patent Document 1, the feedback gain (PI control gain) is changed according to the command voltage. Further, in the motor drive system of Patent Document 2, the feedback gain is changed according to the power fluctuation frequency of the AC motor, and in the voltage converter of Patent Document 3, the feedback gain is changed according to the voltage deviation between the command voltage and the output voltage.

Japanese Patent No. 5206489 JP 2010-124662 A Japanese Patent No. 3969165

  In each of the conventional techniques of Patent Documents 1 to 3, a feedback gain used for PI control is determined based on “one parameter”. According to the applicant's research, when a conventional technology is applied to a voltage converter that uses an inverter that drives a motor generator as a load, the motor generator is adopted as “one parameter” when the motor generator is in a power running operation and a regenerative operation, for example. It has been found that even if the commanded voltage or the rotational speed is the same, the presence or absence of voltage fluctuation is different. Therefore, even if a preferable feedback gain is determined based on one parameter using the prior art, it is not sufficient to suitably reduce the fluctuation of the system voltage depending on all the states of the voltage conversion device and the load. .

  The present invention was created in view of the above points, and an object of the present invention is to select a suitable feedback gain according to the voltage converter and the state of the load, and to reduce the fluctuation of the system voltage. It is to provide.

  A voltage converter according to the present invention is provided between a battery that is a DC power source and a load, and a reactor capable of storing and discharging electric energy, and a high voltage that repeatedly stores and discharges electric energy in the reactor by alternately turning on and off. A potential-side switching element and a low-potential-side switching element, and a controller that calculates a duty command value that is an on-time ratio with respect to a switching cycle of the high-potential-side switching element or the low-potential-side switching element. Convert to system voltage output to load.

  The control unit calculates a feedback term of the duty command value so that the system voltage matches the command voltage, and calculates a feedforward term of the duty command value based on a ratio between the battery voltage and the command voltage A feedforward calculation unit and a gain selection unit are included.

The gain selection unit includes a load fluctuation frequency that is a load fluctuation frequency, a command voltage, a battery current that is a current flowing through the battery, for at least one of a proportional gain and an integral gain among feedback gains used for PI calculation in the feedback calculation unit, The feedback gain is selected based on any two or more specific parameters among five parameters of a load current that is a current flowing through the load and a duty command value.
Here, the specific parameter preferably includes a load fluctuation frequency.

In one aspect of the present invention, the gain selection unit stores a plurality of sets of candidate gains each including a combination of a proportional gain and an integral gain as a matrix, and an evaluation function that indicates a variation amount of the system voltage with respect to two or more sets of candidate gains. It calculates a set of candidate gain for minimizing the evaluation function as the feedback gain.
In another aspect of the present invention, the gain selection unit stores a predetermined basic gain and a plurality of expansion constants capable of performing an extended operation based on a predetermined calculation formula including an operation other than multiplication for the basic gain. An evaluation function indicating the amount of system voltage fluctuation is calculated for a plurality of expansion gains obtained by performing an expansion operation using a plurality of expansion constants on the basic gain, and an expansion gain that minimizes the evaluation function is calculated. Set to feedback gain.
In still another aspect of the present invention, the gain selection unit selects the feedback gain based on any two or more specific parameters among the three parameters of the load fluctuation frequency, the command voltage, and the battery current. In addition, the gain selection unit calculates an evaluation function indicating the variation amount of the system voltage, and selects a feedback gain so as to minimize the evaluation function.

  Since the voltage converter of the present invention selects the feedback gain based on two or more specific parameters, the voltage converter according to the state of the voltage converter and the load is compared with the prior art that determines the feedback gain based on one parameter. More appropriate feedback gain can be selected. Therefore, for example, when an inverter that drives the motor generator is used as a load, fluctuations in the system voltage can be suitably suppressed both during the power running operation and the regenerative operation of the motor generator.

It is a schematic block diagram of the voltage converter by embodiment of this invention. It is a control block diagram of the voltage converter of FIG. 3 is a map of candidate gains used in the first embodiment of the present invention. It is a load fluctuation frequency-evaluation function output characteristic figure in (a) command voltage = Vcom1 and (b) command voltage = Vcom2. (A) A map showing the relationship between battery current and reactor inductance, (b) A map showing the relationship between battery temperature and internal resistance. It is a main flowchart of the duty command value calculation process by 1st Embodiment of this invention. It is a subflowchart of S140 of FIG. It is a map which shows the tendency of the evaluation function output with respect to Kp or Ki. It is a time chart which shows the voltage fluctuation in a comparative example. It is a time chart explaining the voltage fluctuation reduction effect by 1st Embodiment. It is a flowchart of the duty command value calculation process by 2nd Embodiment of this invention. It is a load fluctuation frequency-evaluation function output characteristic figure explaining the gain switching process by 3rd Embodiment of this invention. It is a flowchart of the gain switching process by 3rd Embodiment of this invention. It is a control block diagram of the voltage converter of a comparative example.

Hereinafter, embodiments of a voltage converter according to the present invention will be described with reference to the drawings.
First, a configuration common to a plurality of embodiments will be described. The voltage conversion device 10 of this embodiment is applied to a system that drives an AC motor that is a power source of a hybrid vehicle or an electric vehicle, for example.

[Overall configuration of voltage converter]
An overall configuration of a voltage conversion device according to an embodiment of the present invention will be described with reference to FIG.
As shown in FIG. 1, the voltage conversion device 10 is a device that boosts the DC voltage of the battery 1 and outputs it to an inverter 6 that drives the AC motor 7, and is generally called a “boost converter” or “DCDC converter”. In FIG. 1, a portion surrounded by a one-dot chain line indicates a range of the voltage conversion device 10. The inverter 6 and the AC motor 7 correspond to the “load” of the voltage converter 10.

First, the system configuration outside the range of the voltage converter 10 will be described.
The battery 1 is a direct current power source configured by a chargeable / dischargeable power storage device such as nickel hydride or lithium ion. An electric double layer capacitor or the like is also included in one embodiment of the battery 1.
The battery ECU 55 monitors the power supply voltage VB, battery temperature Tb, SOC (charge amount), etc. of the battery 1 and controls charging / discharging of the battery 1 and the like.

The inverter 6 includes six switching elements that are bridge-connected, and receives DC power of the system voltage Vsys generated by the voltage conversion device 10. The inverter 6 converts the DC power into three-phase AC power and supplies it to the AC motor 7 by turning on and off the switching elements of each phase by PWM control and phase control.
The inverter 6 in FIG. 1 includes an inverter control unit.

  The AC motor 7 is, for example, a permanent magnet type synchronous three-phase AC motor. The AC motor 7 is a motor generator (denoted as “MG” in the figure) mounted on a hybrid vehicle or an electric vehicle, and has a narrow sense that generates torque for driving the drive wheels via a transmission or the like by a power running operation. It has both a function as an electric motor and a function as a generator that generates electric power by a regenerative operation using torque transmitted from the engine and driving wheels. Hereinafter, in this specification, the term “AC motor” is used to mean “motor generator”.

The rotation angle sensor 75 provided in the vicinity of the rotor of the AC motor 7 is constituted by, for example, a resolver or a rotary encoder, and detects the electrical angle θ. The electrical angle θ detected by the rotation angle sensor 75 is input to the inverter 6 and used for calculations such as dq conversion of current vector control. The electrical angle θ is converted into an electrical angular velocity ω by the differentiator 76 and input to the control unit 50.
The electrical angular velocity ω [deg / s] is converted to a rotational speed N [rpm] by multiplying by a proportionality constant. In the present specification and drawings, the “rotational speed converted from the electrical angular velocity ω” is omitted, and the notation “rotational speed ω” is used. Further, “rotational speed ω” corresponds to “load fluctuation frequency f” when the AC motor 7 is generalized as a load. In the description of this embodiment, “rotational speed ω = load fluctuation frequency f” is used.

Next, the configuration of the voltage conversion device 10 will be described. The voltage conversion apparatus 10 includes a filter capacitor 11, a reactor 12, a boost drive unit 15, a smoothing capacitor 16, a control unit 50, and the like.
The filter capacitor 11 is provided at the input unit of the voltage conversion device 10 and removes power supply noise from the battery 1. The reactor 12 has an inductance L, an induced voltage is generated with a change in current, and electric energy is accumulated.

  The step-up drive unit 15 includes a high potential side switching element 13 connected between the output terminal of the reactor 12 and the high potential line of the inverter 6, and between the output terminal of the reactor 12 and the low potential line of the inverter 6. The low-potential side switching element 14 is connected. The high-potential side switching element 13 and the low-potential side switching element 14 are turned on and off alternately and complementarily by a duty command from the control unit 50.

When the high potential side switching element 13 is off and the low potential side switching element 14 is on, the reactor current IL flows through the reactor 12, whereby energy is accumulated.
When the high potential side switching element 13 is on and the low potential side switching element 14 is off, the boosted voltage in which the induced voltage is superimposed on the battery input voltage Vin is released by the smoothing capacitor by releasing the energy stored in the reactor 12. 16 is charged.

On the battery 1 side (input side) of the voltage conversion device 10, the input voltage Vin from the battery 1 is detected. Further, on the inverter 6 side (output side) of the voltage conversion device 10, a system voltage Vsys that is an output DC voltage to the inverter 6 is detected. The system voltage Vsys is also referred to as “inverter input voltage” from the viewpoint of the inverter 6.
The current sensor 17 is provided in the current path from the battery 1 to the voltage conversion device 10 and detects the battery current Ib. The current sensor 18 is provided in the current path from the voltage converter 10 to the inverter 6 and detects the load current Im.

Among the signals input from the outside to the control unit 50 in FIG. 1, the power supply voltage VB, the battery current Ib, the load current Im, and the load fluctuation frequency f (or the rotational speed ω) shown by broken lines are as follows. It is not a signal input in common to all the embodiments, but is input as necessary according to the configuration of each form.
For example, the current sensor 17 may be omitted when the control unit 50 does not use the battery current Ib information, and the current sensor 18 may be omitted when the load current Im information is not used. Further, when the control unit 50 does not store the battery temperature-internal resistance map shown in FIG. 5B, the battery temperature Tb may not be input from the battery ECU 55. Which input signal information the control unit 50 uses is determined by selecting a specific parameter, which will be described later.

  Next, the configuration of the control unit 50 will be described with reference to FIG. The control unit 50 is configured by a microcomputer or the like, and includes a CPU, a ROM, an I / O (not shown), a bus line that connects these configurations, and the like. The control unit 50 executes control by software processing by executing a program stored in advance by the CPU or by hardware processing by a dedicated electronic circuit.

The control unit 50 includes a command voltage generation unit 29, a feedback calculation unit 20, a feedforward calculation unit 30, and a gain selection unit 25 that is a characteristic configuration of the present invention.
The control unit 50 sets the duty command value duty for defining the switching time of the high-potential side switching element 13 and the low-potential side switching element 14 so that the system voltage Vsys generated by the boost drive unit 15 matches the command voltage Vcom. To control the drive of the step-up drive unit 15.

In the present embodiment, the on-duty indicating the on-time ratio with respect to the switching period of the high potential side switching element 13 is defined as “duty command value duty (%)”. If the dead time is ignored, the on-duty of the low-potential side switching element 14 coincides with the off-duty of the high-potential side switching element 13 and corresponds to “100−duty (%)”.
In another embodiment, the on-duty of the low potential side switching element 14 may be defined as “duty command value duty (%)”.

The command voltage generator 29 calculates the command voltage Vcom based on the command torque trq ^* for the AC motor 7 commanded from the host vehicle control circuit and the rotational speed ω of the AC motor 7 acquired from the rotation angle sensor 75. .
In another embodiment, the control unit 50 may acquire the command voltage Vcom calculated by the host vehicle control circuit instead of including the command voltage generation unit 29 inside the control unit 50.

  The feedback calculation unit 20 includes a proportional gain calculator 21 that calculates a proportional term, and an integral gain calculator 22 and an integrator 23 that calculate an integral term. The feedback calculator 20 receives a deviation between the command voltage Vcom and the system voltage Vsys and calculates a feedback term dfb of the duty command value duty by PI calculation so that the deviation converges to zero.

  Here, the proportional gain Kp used by the proportional gain calculator 21 and the integral gain Ki used by the integral gain calculator 22 are collectively referred to as “feedback gain”. In this embodiment, the feedback calculation unit 20 does not use the proportional gain Kp and the integral gain Ki as fixed values, but uses the proportional gain Kp and the integral gain Ki selected by the gain selection unit 25 at that time. .

The feedforward calculation unit 30 calculates the feedforward term ff of the duty command value duty based on the ratio between the input voltage Vin and the command voltage Vcom. Instead of the input voltage Vin, the power supply voltage VB input from the battery ECU 55 may be used.
The controller 50 outputs a duty command value duty obtained by adding the feedback term dfb and the feedforward term ff to the boost drive unit 15.
Further, the command voltage Vcom and the duty command value duty generated inside the control unit 50 are input to the gain selection unit 25 as necessary.

[Configuration of gain selector]
As indicated by a broken line in FIG. 2, the gain selection unit 25 includes a load fluctuation frequency f (rotational speed ω), a battery current Ib, a load current Im, and the control unit 50 that are input from the outside of the control unit 50. Any two or more parameters among the five parameters of the generated command voltage Vcom and duty command value duty are input. Then, the feedback gain is selected based on the two or more input parameters. Here, a reactor current may be used instead of the battery current Ib.
Hereinafter, two or more parameters used as information for feedback gain selection by the gain selection unit 25 are referred to as “specific parameters”. A specific method by which the gain selection unit 25 selects the feedback gain based on the specific parameter will be described for each embodiment.

(First embodiment)
The configuration and effect of the feedback gain selection process according to the first embodiment will be described with reference to FIGS.
As shown in FIG. 3, the gain selection unit 25 matrixes combinations of candidate values of the proportional gain Kp and the integral gain Ki that the feedback calculation unit 20 may use for PI control (hereinafter referred to as “candidate gain”). Is stored in the map. FIG. 3 exemplifies a matrix of n rows and m columns for the proportional gain Kp and n for the integral gain Ki.

The gain selection unit 25 selects an optimum combination of feedback gains from map candidate gains. Here, “optimum” is not limited to the best one in a strict sense, but is interpreted to mean that it is included in an “optimum group” that is particularly good compared to most other candidate gains.
As the selection method, the gain selection unit 25 calculates an “evaluation function” corresponding to the candidate gain based on two or more specific parameters, and sets the candidate gain that minimizes the evaluation function as a feedback gain.

  In the present embodiment, the “evaluation function” is defined as “an index indicating the amount of change in the system voltage Vsys according to the state of the voltage converter and the load”. The evaluation function is a function including, for example, the above five parameters as arguments, and a specific calculation formula is appropriately designed by analyzing system characteristics. Since the evaluation function used in the present embodiment indicates the amount of fluctuation of the system voltage Vsys, it is determined that the load can be driven more stably as the output value is smaller.

  FIG. 4 shows an example of an evaluation function when the load fluctuation frequency f and the command voltage Vcom are used as specific parameters. 4A is a characteristic diagram of the load fluctuation frequency f and the evaluation function output when the command voltage Vcom is Vcom1 which is relatively high, and FIG. 4B is a characteristic diagram when the command voltage Vcom is Vcom2 which is relatively low. .

In the figure, nine thin lines (solid line, one-dot chain line, and broken line) indicate evaluation function outputs calculated for nine candidate gains. The evaluation function corresponding to each gain has a substantially mountain shape having one maximum value in the middle of the frequency, and the shape, height, frequency of the maximum value, and the like are different from each other.
When the nine evaluation function outputs are superimposed and plotted, and attention is paid to the minimum value, the gain that minimizes the evaluation function alternates at a certain switching frequency as indicated by a bold line. Unlike the switching frequency fx1 when the command voltage is Vcom1 and the switching frequency fx2 when the command voltage is Vcom2, in this example, fx1 <fx2. In the region below the switching frequency fx1 or fx2, the evaluation function E (G1) corresponding to the candidate gain G1 is minimized, and in the region exceeding the switching frequency fx1 or fx2, the evaluation function E (G9) corresponding to the candidate gain G9 is obtained. Minimal.

As can be seen from FIG. 4, when the command voltage Vcom is constant, the candidate gain to be selected differs depending on the load fluctuation frequency f. When the load fluctuation frequency f is fy between fx1 and fx2, if the command voltage is Vcom1, the candidate gain G9 is selected, and if the command voltage is Vcom2, the candidate gain G1 is selected.
Thus, in this embodiment, the candidate gain that minimizes the evaluation function changes according to the two specific parameters of the load fluctuation frequency f and the command voltage Vcom.

The specific parameters used for feedback gain selection are not limited to the load fluctuation frequency f and the command voltage Vcom, but any two or more of the five parameters including the battery current Ib, the load current Im, and the duty command value duty are selected. May be. However, since the influence of the frequency response on the evaluation function is particularly large compared to other parameters, it is preferable to include the load fluctuation frequency f as one of the specific parameters.
In addition, as the number of specific parameters increases, the calculation accuracy of the evaluation function improves, but the processing load and the required storage area increase. Therefore, the balance between the calculation accuracy required for the system and the processing capability of the controller 50 The number of specific parameters may be determined in consideration of

  In addition to the above five parameters, numerical information that becomes an argument of the evaluation function includes an inductance value of the reactor 12, an internal resistance value of the battery 1, and the like. For simplicity, the control unit 50 may store these numerical values as fixed values and use them for calculation.

However, preferably, the control unit 50 may store a battery current-reactor inductance characteristic map shown in FIG. 5A and a battery temperature-internal resistance map shown in FIG. And it is good to change the inductance value of the reactor 12 used for calculation of an evaluation function according to the battery current Ib, and to change the internal resistance value of the battery 1 used for calculation of the evaluation function according to the battery temperature Tb. Thereby, the calculation accuracy of the evaluation function can be improved.
5A and 5B is not limited to the form performed by the gain selection unit 25 itself, and is stored and calculated in another control block of the control unit 50, and the calculation referring to the map is performed. The result may be output to the gain selection unit 25.

Next, the duty command value calculation process of the first embodiment will be described with reference to the flowcharts of FIGS. 6 and 7 and the map of FIG. In the description of the flowchart below, the symbol “S” means a step.
The purpose of this flow is to efficiently search for a candidate gain that minimizes the evaluation function from among the candidate gains of the matrix-like map shown in FIG. 3, thereby calculating an appropriate duty command value duty. To do.

In S11 of FIG. 6, the gain selection unit 25 selects one set of candidate gains from the matrix-like map, and calculates an evaluation function corresponding to the selected candidate gains.
In S12, the value of the evaluation function calculated this time is compared with “the minimum value of the evaluation function calculated so far”. When the value of the evaluation function calculated this time is smaller than the minimum value of the evaluation function calculated so far (S12: YES), the minimum value of the evaluation function is updated in S13. In the case of NO, the minimum value of the evaluation function so far is maintained.

  According to the technical idea of the present invention, theoretically, an evaluation function may be calculated for all candidate gains of the matrix of n rows and m columns, and the minimum values may be compared. However, in reality, if the evaluation function is calculated for all candidate gains, a large costly circuit is required due to an increase in processing load and an increase in the storage area for candidate gains, and the calculation cycle becomes long. There is. In addition, it is inefficient to calculate the evaluation function for all candidate gains that are estimated to be clearly larger than the minimum value until then.

Therefore, in S140, candidate gains that need to be evaluated are selected from the tendency of the evaluation functions calculated so far. In other words, candidate gains that are determined to be unnecessary to calculate the evaluation function are excluded, and the candidate gains are narrowed down.
Specifically, for example, candidate gains are selected by the subflow shown in FIG.

  In S141 of FIG. 7, when the individual relationship between the proportional gain Kp or the integral gain Ki and the evaluation function is arranged for the candidate gains calculated so far, the monotonic increase as shown in FIGS. 8A and 8B is performed. Alternatively, it is determined whether a monotonous decrease relationship is established. If YES in S141, it is assumed that the evaluation function for the candidate gain including the intermediate value of the proportional gain Kp or the integral gain Ki whose relationship has been examined should not correspond to the maximum value or the minimum value, and the evaluation function in S142 The operation of is omitted.

  Assume that the proportional gains Kp1... Kpn and the integral gains Ki1... Kim in the matrix of FIG. 3 are arranged in order from a small value to a large value. Are monotonically increasing or monotonically decreasing, it is determined that the evaluation function only needs to be calculated for the candidate gains at the four corners indicated by the broken line and the alternate long and short dash line. Further, when the increasing or decreasing tendency with respect to the proportional gain Kp and the integral gain Ki coincides, it is determined that it is only necessary to calculate the evaluation function for the two upper left and lower right candidate gains indicated by the broken lines. For example, also in FIG. 4, the candidate gains constituting the minimum value are only G1 and G9 at both ends of the nine, and the seven intermediate candidate gains are not used after all even if the evaluation function is calculated.

  In S143, for the candidate gains calculated so far, as shown in FIG. 8C, whether the change in the evaluation function with respect to the change in the proportional gain Kp or the integral gain Ki tends to be saturated, in other words, the evaluation function Is asymptotic to the saturation value Esat. In the case of YES in S143, it is estimated that the evaluation function does not change even if the evaluation function is further increased or decreased in the direction in which the evaluation function is saturated with respect to the proportional gain Kp or the integral gain Ki for which the change tendency is examined. Stop increasing or decreasing.

  As described above, in S140, the control unit selects the candidate gain that can omit the calculation of the evaluation function and the candidate gain that needs to be calculated thereafter, while monitoring the tendency of the evaluation function calculated so far. The storage area for 50 processing loads and candidate gains can be reduced.

  However, when an extreme value exists in the relationship between the proportional gain Kp or the integral gain Ki and the evaluation function, or the evaluation function changes irregularly with respect to the proportional gain Kp or the integral gain Ki, the evaluation function There may be few candidate gains that can be omitted.

In S15, it is determined whether or not all of the “candidate gains that require calculation of the evaluation function” selected in S140 are completed. If not, the process returns to S11.
In S16, the candidate gain corresponding to the minimum value of the evaluation function is set as the feedback gain. The feedback calculation unit 20 performs a feedback calculation using the set proportional gain Kp and integral gain Ki (S17). The control unit 50 calculates the duty command value duty by adding the calculation result of the feedback calculation and the calculation result of the feedforward calculation (S18) by the feedforward calculation unit 30 (S19).

  Subsequently, the effects of the present embodiment will be described with reference to FIGS. 9 and 10 while being compared with the comparative example. In the time charts shown in FIGS. 9 and 10 (a) to (c), the common time axis is the horizontal axis, (a) accelerator opening, (b) rotational speed ω of AC motor 7, (c) system. The voltage Vsys is the vertical axis. FIG. 9 shows the characteristics according to the comparative example, and FIG. 10 shows the characteristics according to the present embodiment. 9 and 10, (a) and (b) are common and only (c) is different.

As shown in FIG. 14, the control unit 90 of the voltage conversion device of the comparative example is similar to the present embodiment in that the feedback calculation unit 20 includes a gain selection unit 95 that selects a feedback gain used for PI control. . However, the point which the gain selection part 95 selects a feedback gain based on only one parameter differs from the gain selection part 25 of this embodiment.
For example, in the comparative example, the feedback gain is selected based on only one of the rotational speed ω (= load fluctuation frequency f) or the command voltage Vcom, whereas in the present embodiment, the feedback gain is selected based on both the rotational speed ω and the command voltage Vcom. The feedback gain is selected.

9 and 10, the initial value of the system voltage Vsys at time t0 is α.
When the accelerator opening starts to increase at time t0, AC motor 7 is accelerated by the power running operation while consuming DC power of battery 1, and rotational speed ω increases. In response to this, the command voltage generator 29 of the voltage converter 10 generates a command voltage Vcom corresponding to the target value β in order to change the system voltage Vsys to a target value β that is larger than the initial value α.
The feedback calculation unit 20 calculates the feedback term dfb by PI control that attempts to make the deviation between the command voltage Vcom and the system voltage Vsys zero, and the feedforward calculation unit 30 calculates the ratio between the battery input voltage Vin and the command voltage Vcom. Based on the above, the feedforward term ff is calculated. Thus, the system voltage Vsys increases from the initial value α toward the target value β. The rotational speed ω at time t1 immediately after the system voltage Vsys reaches the target value β is ωa.

Thereafter, when the accelerator opening degree starts to decrease at time t2, AC motor 7 switches from acceleration to deceleration, and regenerates electric power generated by the rotation in battery 1.
In response to the decrease in the rotational speed ω, the command voltage generation unit 29 of the voltage conversion device 10 changes the system voltage Vsys from the current value β to the target value α equivalent to the initial value, so as to change the command voltage corresponding to the target value α. Generate Vcom. The system voltage Vsys decreases from the current value β toward the target value α immediately after time t3 when the rotational speed ω decreases to ωa, for example, by the control of the feedback calculation unit 20 and the feedforward calculation unit 30 that follow this. . At time t4, the system voltage Vsys returns to the target value α equivalent to the initial value.

  In the transition of the system voltage Vsys as described above, as shown in FIG. 9, in the comparative example, voltage fluctuation occurs near time t1 and immediately after time t4. Here, at time t1 and time t3, although the rotational speed ω is equal, voltage fluctuation occurs at time t1 during the power running operation, and voltage fluctuation does not occur at time t3 during the regenerative operation. . Further, at time t0 and time t4, although the command voltage Vcom is equivalent, no voltage fluctuation occurs at time t0 during the power running operation, and voltage fluctuation occurs at time t4 during the regenerative operation. .

That is, even if the rotation speed ω or the command voltage Vcom is the same, the states of the voltage conversion device 10 and the loads 6 and 7 are not exactly the same. Therefore, when the feedback gain is selected based on only one of the rotation speed ω and the command voltage Vcom, the occurrence of voltage fluctuation cannot be sufficiently suppressed.
On the other hand, as shown in FIG. 10, in this embodiment, it turns out that the voltage fluctuation has not generate | occur | produced in any time when the voltage fluctuation has generate | occur | produced in the comparative example. Therefore, in the present embodiment in which the feedback gain is selected based on both the rotational speed ω and the command voltage Vcom, fluctuations in the system voltage Vsys can be suitably suppressed as compared with the comparative example.

(Second Embodiment)
The duty command value calculation processing according to the second embodiment of the present invention will be described with reference to the flowchart of FIG. In FIG. 11, steps that are substantially the same as those in FIG. 6 of the first embodiment are denoted by the same step numbers and description thereof is omitted.
In the second embodiment, instead of the “candidate gain” in the first embodiment, a predetermined “basic gain” and an “expansion gain” obtained by multiplying the basic gain by an expansion constant, addition, and the like are used as feedback gains. To be selected. Accordingly, S11, S140, S15, and S16 in FIG. 6 are changed to S21, S240, S25, and S26 in FIG.

  The basic gain G0 is assumed to be a combination of a basic proportional gain Kp0 and a basic integral gain Ki0. The expansion constant C is a combination of a proportional gain expansion constant Cp and an integral gain expansion constant Ci. The proportional gain expansion constant Cp and the integral gain expansion constant Ci may be set to different values, but here, for simplicity, “C = Cp = Ci”. The gain selection unit 25 stores a plurality of expansion constants C, that is, C1, C2,... Cn.

  Further, a predetermined calculation formula that is implemented using the expansion constant C for the basic gain G0 is defined. For example, in the case of multiplication, the expansion gain G ′ is calculated by the equation “G ′ = C × G0”. In this case, the extended gain G ′ means a combination of a proportional gain (C × Kp0) and an integral gain (C × Ki0). Alternatively, when the predetermined calculation formula is addition, the expansion gain G ′ is calculated by the formula “G ′ = G0 + C”. This multiplication and addition is collectively referred to as “extended operation”.

  If the number of the plurality of expansion constants C1, C2,... Cn is n, n expansion gains G1 ′, G2 ′,. The expansion constant C is, for example, “2” or “0.5”. When the expansion constant C includes “1”, the basic gain G0 itself is also included in the expansion gain G ′.

In S21 of FIG. 11, one of a plurality of expansion constants C is selected, the expansion gain G ′ is calculated by performing expansion operations such as multiplication and addition on the basic gain G0, and the obtained expansion gain G The evaluation function corresponding to 'is calculated.
In S12 and S13, as in the first embodiment, if the evaluation function calculated this time is smaller than the minimum value of the evaluation functions calculated so far, the minimum value is updated.

  S240 corresponds to S140 of the first embodiment, and it is necessary to calculate an evaluation function for the expansion gain G ′ calculated using the expansion constant C stored by the gain selection unit 25. Sort out what is missing. About a specific subflow, since only a term is replaced with FIG. 7 of 1st Embodiment, description is abbreviate | omitted.

In S25, it is determined whether or not all of the “extended gains that require calculation of the evaluation function” selected in S240 are completed. If not, the process returns to S21.
In S26, the expansion gain corresponding to the minimum value of the evaluation function is set as the feedback gain. The steps after S17 are the same as those in the first embodiment.

  In the second embodiment, in addition to the same effects as the first embodiment, a plurality of candidate gains are stored in a matrix-like map, and the evaluation function is calculated by selecting candidate gains from this map. The processing load and the storage area can be reduced by storing a simpler expansion constant C for the form and using it for the calculation.

(Third embodiment)
A third embodiment of the present invention will be described with reference to FIGS. In the first and second embodiments described above, the evaluation function is calculated based on the specific parameter at the time of executing the process, and the optimum feedback gain is selected at that time. Is focused on.

  On the other hand, in the third embodiment, the feedback calculation is performed when the selected feedback gain is changed as a result of the change of the specific parameter during the execution of the feedback calculation using the feedback gain selected in the first or second embodiment. A preferable “gain switching process” for switching the feedback gain used by the unit 20 is provided. In other words, it focuses on the transition from point to point on the time axis, and may be combined with any of the first and second embodiments.

FIG. 12 is a diagram showing a load fluctuation frequency f-evaluation function output characteristic similar to FIG. The evaluation function E (G1) corresponding to the candidate gain G1 is the smallest in the region where the load fluctuation frequency f is less than fx, and the evaluation function E (G2) corresponding to the candidate gain G2 is the region where the load fluctuation frequency f exceeds fx. Is the smallest. Therefore, simply, when the load fluctuation frequency f changes across fx, it is considered that the feedback gain may be switched.
However, in this case, when the load fluctuation frequency repeatedly increases and decreases across fx, a hunting phenomenon in which the feedback gain frequently switches may occur, and control may become unstable.

Therefore, as indicated by thick broken lines in FIG. 12, characteristic lines E (G1) ′ and E (G2) offset by a predetermined value ε (> 0) on the negative side with respect to the evaluation functions E (G1) and E (G2). Assuming ', in the state before switching, the value of the offset characteristic line is regarded as the value of the evaluation function.
When the load fluctuation frequency f changes in the increasing direction, the feedback gain is switched from G1 to G2 at the first switching value fxH where E (G1) ′ and E (G2) intersect, and the load fluctuation frequency f changes in the decreasing direction. In this case, the feedback gain is switched from G2 to G1 at the second switching value fxL where E (G2) ′ and E (G1) intersect. Since the predetermined value ε is set on the minus side, the first switching value fxH is inevitably set to a value larger than the second switching value fxL, that is, on the high frequency side.

Next, in the gain switching process of the third embodiment shown in the flowchart of FIG. 13, first, in S31, it is determined whether the gain G1 is currently selected or the gain G2 is being applied.
While the gain G1 is being applied (S31: YES), the load fluctuation frequency f is increasing (S32: YES), and when it becomes equal to or higher than the first switching value fxH (S33: YES), the gain G2 is switched (S34). On the other hand, when the gain G2 is being applied (S31: NO), the load fluctuation frequency f is decreasing (S35: YES), and when it becomes equal to or less than the second switching value fxL (S36: YES), the gain G1 is switched (S37). ).

As described above, in the third embodiment, by using the hysteresis characteristic, it is possible to prevent hunting accompanying switching of the feedback gain. Moreover, the range which prevents hunting can be easily changed by adjusting predetermined value (epsilon).
12 and 13 exemplify the case where one of the specific parameters is the load fluctuation frequency f, but the case where the gain that minimizes the evaluation function is switched according to the increase or decrease of the other specific parameters can be handled in the same manner. it can.

(Other embodiments)
(A) In the above embodiment, the “evaluation function” is defined as “an index indicating the amount of fluctuation of the system voltage Vsys according to the voltage converter and the load state”, and the candidate gain or the extension gain that minimizes the evaluation function is defined as Although the feedback gain is set, the definition of the evaluation function is not limited to this. For example, an evaluation function reflecting “load stability” may be set, and a feedback gain that maximizes the evaluation function may be selected.

(A) Sub-flowchart of the first embodiment In FIG. 7, while monitoring the tendency of the evaluation function calculated so far, candidate gains that can omit the calculation of the evaluation function are selected, and the calculation amount of the evaluation function is reduced. ing. In addition, the relationship between the voltage conversion device and the load in the past and the feedback gain selected in that state may be learned, and the feedback gain selection process may be omitted to the minimum based on the learning result.
On the contrary, when the processing capability of the control unit 50 is sufficient, the step of S140 in FIG. 6 may be omitted, and the evaluation function may always be calculated for all combinations of candidate gains in the matrix. This also applies to S240 of FIG. 11 in the second embodiment.

(C) In the second embodiment, the example in which the expansion gain G ′ is calculated by multiplying or adding the expansion constant C to the basic gain G0 has been described. Other operations such as exponential calculation (G0 ^C ), logarithmic calculation (log _C G0), and trigonometric function calculation (tan (G0 / C)) using the expansion constant C for the basic gain G0 Then, the expansion gain G ′ may be calculated.

(D) In the above embodiment, the gain selection unit 25 selects the optimum feedback gain for both the proportional gain Kp and the integral gain Ki based on the specific parameter. However, in the present invention, only one of the proportional gain Kp and the integral gain Ki may be selected based on the specific parameter, and the other gain may be determined by applying the conventional technique, or may be a fixed value.
When the feedback calculation unit performs PID control including differential control, the feedback gain may be selected by applying the present invention to the differential gain Kd.

(E) The voltage converter of the present invention is not limited to a boost converter that boosts and outputs an input side voltage, but may be a step-down converter that steps down and outputs an input side voltage.
(F) In the above embodiment, examples of using the inverter 6 that converts a DC voltage into a three-phase AC voltage and the AC motor 7 that is driven by the three-phase AC voltage converted by the inverter 6 are used as the load of the voltage converter 10. Explained. In addition, as a load of the voltage conversion apparatus 10, for example, an H bridge circuit and a DC motor may be used.

(G) AC motors as loads are not limited to those used as power sources for hybrid vehicles and electric vehicles, but are used for vehicle accessories, trains other than vehicles, elevators, general machinery, etc. May be. In particular, the voltage converter of the present invention is effectively applied to a load having a large change in operating state.
As mentioned above, this invention is not limited to the said embodiment at all, In the range which does not deviate from the meaning of invention, it can implement with a various form.

10: Voltage converter,
12 ... Reactor,
13 ... high potential side switching element,
14 ... low potential side switching element,
20 ... Feedback calculation unit,
25 ... Gain selection section,
30: Feedforward calculation unit,
50 ... control unit,
6 ... Inverter (load),
7: AC motor (motor generator, load).

Claims (8)

Provided between the battery (1) which is a DC power source and the load (6, 7),
A reactor (12) capable of storing and releasing electrical energy;
A high-potential-side switching element (13) and a low-potential-side switching element (14) that repeatedly accumulate and release electric energy in the reactor by alternately turning on and off;
A control unit (50) for calculating a duty command value (duty) which is an on-time ratio with respect to a switching cycle of the high potential side switching element or the low potential side switching element, and an input voltage (Vin) from the battery Is converted to a system voltage (Vsys) to be output to the load,
The controller is
A feedback calculation unit (20) for calculating a feedback term of the duty command value so that the system voltage matches the command voltage (Vcom);
A feedforward calculation unit (30) for calculating a feedforward term of the duty command value based on a ratio between the battery voltage (Vin, VB) and the command voltage;
A battery that is a load fluctuation frequency (f) that is a fluctuation frequency of the load, a command voltage, and a current that flows through the battery for at least one of a proportional gain and an integral gain among feedback gains used for PI calculation in the feedback calculation unit. A gain selection unit that selects a feedback gain based on any two or more of the five parameters of the current (Ib), the load current (Im) that is the current flowing through the load, and the duty command value; 25) ,
The gain selection unit stores a plurality of sets of candidate gains each including a combination of a proportional gain and an integral gain as a matrix, and calculates an evaluation function indicating a variation amount of the system voltage for the two or more sets of candidate gains. The voltage conversion apparatus characterized in that the candidate gain that minimizes the evaluation function is set as a feedback gain . Provided between the battery (1) which is a DC power source and the load (6, 7),
A reactor (12) capable of storing and releasing electrical energy;
A high-potential-side switching element (13) and a low-potential-side switching element (14) that repeatedly accumulate and release electric energy in the reactor by alternately turning on and off;
A control unit (50) for calculating a duty command value (duty) which is an on-time ratio with respect to a switching cycle of the high potential side switching element or the low potential side switching element, and an input voltage (Vin) from the battery Is converted to a system voltage (Vsys) to be output to the load,
The controller is
A feedback calculation unit (20) for calculating a feedback term of the duty command value so that the system voltage matches the command voltage (Vcom);
A feedforward calculation unit (30) for calculating a feedforward term of the duty command value based on a ratio between the battery voltage (Vin, VB) and the command voltage;
A battery that is a load fluctuation frequency (f) that is a fluctuation frequency of the load, a command voltage, and a current that flows through the battery for at least one of a proportional gain and an integral gain among feedback gains used for PI calculation in the feedback calculation unit. A gain selection unit that selects a feedback gain based on any two or more of the five parameters of the current (Ib), the load current (Im) that is the current flowing through the load, and the duty command value; 25) ,
The gain selection unit stores a predetermined basic gain and a plurality of expansion constants capable of performing an extended operation based on a predetermined calculation formula including an operation other than multiplication for the basic gain, and the basic gain For the plurality of expansion gains obtained by performing the expansion calculation using a plurality of the expansion constants, an evaluation function indicating the variation amount of the system voltage is calculated, and the expansion gain that minimizes the evaluation function is calculated. A voltage conversion device characterized by being set to a feedback gain . Provided between the battery (1) which is a DC power source and the load (6, 7),
A reactor (12) capable of storing and releasing electrical energy;
A high-potential-side switching element (13) and a low-potential-side switching element (14) that repeatedly accumulate and release electric energy in the reactor by alternately turning on and off;
A control unit (50) for calculating a duty command value (duty) which is an on-time ratio with respect to a switching cycle of the high potential side switching element or the low potential side switching element, and an input voltage (Vin) from the battery Is converted to a system voltage (Vsys) to be output to the load,
The controller is
A feedback calculation unit (20) for calculating a feedback term of the duty command value so that the system voltage matches the command voltage (Vcom);
A feedforward calculation unit (30) for calculating a feedforward term of the duty command value based on a ratio between the battery voltage (Vin, VB) and the command voltage;
A battery that is a load fluctuation frequency (f) that is a fluctuation frequency of the load, a command voltage, and a current that flows through the battery for at least one of a proportional gain and an integral gain among feedback gains used for PI calculation in the feedback calculation unit. current (Ib), gain selection unit that selects the feedback gain based on any two or more specific parameters of the three parameters (25) has a,
The gain selection unit calculates an evaluation function indicating a variation amount of the system voltage, and selects a feedback gain so as to minimize the evaluation function . In the process of calculating the evaluation function by sequentially using a plurality of feedback gains and updating the minimum value of the evaluation function, the gain selection unit needs to calculate from the tendency of the evaluation function calculated so far. The voltage converter as described in any one of Claims 1-3 which select a feedback gain. The specific parameter is a voltage conversion device according to claim 1, characterized in that it comprises the load variation frequency. In the case where the value of the specific parameter changes and the feedback gain selected by the gain selection unit is switched,
The first switching value (fxH) for switching the feedback gain when the specific parameter changes in the increasing direction is set to a value larger than the second switching value (fxL) for switching the feedback gain when the specific parameter changes in the decreasing direction. The voltage converter according to any one of claims 1 to 5, wherein The voltage conversion device according to any one of claims 1 to 6, wherein the control unit changes an inductance value of the reactor used for calculation of the evaluation function in accordance with the battery current. The voltage conversion device according to any one of claims 1 to 7, wherein the control unit changes an internal resistance value of the battery used for calculation of the evaluation function in accordance with a temperature of the battery. .

Images