JP2018074751A - Controller for dc-dc converter - Google Patents

Abstract

PROBLEM TO BE SOLVED: To achieve stable control of a DC-DC converter.SOLUTION: A controller 104 includes: a MPC which predicts a predetermined state value at prediction time in a power converter 100, using a state equation of the power converter 100 and controls the power converter 100 according to the predicted state value; a prediction time calculator which sets prediction time in the MPC.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a control device for a DC / DC converter.

FIG. 25 is a diagram showing a configuration of a conventional step-up DC / DC converter. In a step-up DC / DC converter, switching control of a switching element such as an FET is performed to boost the voltage of the power source E and output it to the load R as the output voltage _vo .

A control device in which a model predictive controller MPC is applied to such a step-up DC / DC converter as shown in FIG. 26 is disclosed (Non-Patent Document 1). A command value v _r ′ is generated from an error between the target value v _r and the output voltage v _o with respect to the output voltage, and the model predictive controller MPC according to the command value v _r ′ and the current state value x of the DC / DC converter. The state of the DC / DC converter is predicted, and the DC / DC converter is controlled based on the predicted value.

“Model Predictive Control of Boost DC-DC Converter and its Implementation”, Liu Yasushi, Yasushi Katane, 2013 IEEJ National Convention 4-003

  By the way, the step-up DC / DC converter targeted by the conventional control technique has a configuration in which there is one switching element and no reflux diode. On the other hand, in the DC / DC converter in which the reactor is connected to the connection point between the first switching element constituting the upper arm and the second switching element constituting the lower arm, a technique for causing the reactor current to follow the reactor current command value. Is applied to model predictive control that causes the capacitor voltage to follow the capacitor voltage command value, the duty ratio of the switching element tends to oscillate and cannot be controlled stably.

  One aspect of the present invention is a control device that controls a DC / DC converter, predicting a predetermined state value in a prediction time in the DC / DC converter using a state equation of the DC / DC converter, A control apparatus comprising: a model prediction controller that controls the DC / DC converter according to the predicted state value; and a prediction time calculator that sets the prediction time in the model prediction controller. It is.

  Here, it is preferable that the prediction time calculator sets the prediction time using a reactor value and a capacitor value included in the DC / DC converter.

The DC / DC converter includes a reactor connected to a power source, a first switching element and a second switching element that control a current flowing through the reactor, and a capacitor that smoothes an output voltage from the reactor. wherein the model predictive controller, the power supply voltage v _b of the power source, said predetermined condition by applying the output current i _m of the capacitor voltage v _c and the DC / DC converter at both ends of the capacitor to the state equation A control device that predicts a value and controls the DC / DC converter according to the predicted state value.

Here, an observer is provided that estimates an error duty ratio Δd with respect to a duty ratio d that is a ratio of an on period of the first switching element and the second switching element according to a current state value of the DC / DC converter, model predictive controller, wherein the power supply voltage v _b, in addition to said capacitor voltage v _c, the output current i _m, predicting the predetermined state value by applying said error duty ratio Δd in the state equation Is preferred.

Further, an error duty ratio Δd with respect to a duty ratio d, which is a ratio of an ON period of the first switching element and the second switching element, according to a current state value of the DC / DC converter, and an estimation of a reactor current flowing through the reactor It comprises an observer that estimates the value i _L ^~, the model predictive controller, wherein the power supply voltage v _b, the capacitor voltage v _c, in addition to the output current i _m, the error duty ratio Δd and the reactor current estimated value It is preferable to predict the predetermined state value by applying i _L ^{to the} state equation.

  Moreover, it is preferable to set the value of the reactor according to the current flowing through the reactor, and to apply the value of the reactor to the state equation.

  According to the present invention, the DC / DC converter can be stably controlled by appropriately setting the prediction time for predicting the state value and the settling time for setting the command value in the control for the DC / DC converter. it can.

It is a figure which shows the basic composition of the power converter device in 1st Embodiment. It is a figure which shows the structure of the control apparatus in 1st Embodiment. It is a figure which shows the relationship between a reactor current and the value (inductance) of a reactor. It is a figure which shows the structure of the model prediction controller in 1st Embodiment. It is a figure which shows the output characteristic of a power converter device. It is a figure which shows the output characteristic of a power converter device. It is a graph which shows the example of the voltage control in 1st Embodiment. It is a graph which shows the example of the current control in 1st Embodiment. It is a graph which shows the time change of the solution of the quadratic equation in a 1st embodiment. It is a graph which shows the example of the voltage control in a comparative example. It is a graph which shows the example of the current control in a comparative example. It is a graph which shows the time change of the solution of the quadratic equation in a comparative example. It is a figure which shows the basic composition of the power converter device in 2nd Embodiment. It is a figure which shows the structure of the control apparatus in 2nd Embodiment. It is a figure which shows the structure of the model prediction controller in 2nd Embodiment. It is a figure which shows the basic composition of the power converter device in 4th Embodiment. It is a figure which shows the structure of the control apparatus in 4th Embodiment. It is a figure which shows the structure of the model prediction controller in 4th Embodiment. It is a graph which shows the example of the voltage control in 4th Embodiment. It is a graph which shows the example of the current control in 4th Embodiment. It is a graph which shows the time change of the solution of the quadratic equation in 4th Embodiment. It is a graph which shows the example of the voltage control in a comparative example. It is a graph which shows the example of the current control in a comparative example. It is a graph which shows the time change of the solution of the quadratic equation in a comparative example. It is a figure which shows the structure of the conventional power converter device. It is a figure which shows the structure of the control apparatus of the power converter device by a prior art.

<First Embodiment>
FIG. 1 shows a basic configuration of a power conversion device 100 according to an embodiment of the present invention. The power conversion apparatus 100 includes a DC power supply 10, a reactor 12, a first switching element 14, a second switching element 16, and a capacitor 18.

  One end of the reactor 12 is connected to the positive electrode of the DC power supply 10, and the connection point C of the first switching element 14 and the second switching element 16 is connected to the other end of the reactor 12. The other end of the first switching element 14 is connected to the output terminal (OUT +) to the load 102, and the other end of the second switching element 16 is connected to the negative electrode (OUT−) of the DC power supply 10. A capacitor 18 is connected between the output terminal and the negative electrode of the DC power supply 10 in order to smooth the voltage.

  In the present embodiment, the first switching element 14 and the second switching element 16 are NPN transistors. The first switching element 14 has a collector on the output end side and an emitter on the reactor 12 side. The second switching element 16 has a reactor 12 side as a collector and a DC power supply 10 negative electrode side as an emitter. A free-wheeling diode is connected in parallel to each of the first switching element 14 and the second switching element 16.

In the power conversion device 100, the reactor current i _L flows from the positive electrode to the negative electrode of the DC power supply 10 through the reactor 12 by turning the first switching element 14 off and the second switching element 16 on. As a result, energy is accumulated in the reactor 12. Next, the first switching element 14 is turned on and the second switching element 16 is turned off, whereby the reactor current i _L is cut off, and the voltage of the DC power source 10 (battery voltage v _b ) is connected to the end of the reactor 12. cause a voltage higher than, current corresponding to the capacitor voltage v _c rises capacitor 18 flows is charged towards the output end. The capacitor voltage v _c is applied to the load 102. Output voltage, i.e. the capacitor voltage v _c of the power converter 100 is determined by the duty ratio which is the ratio of the on period of the first switching element 14 and second switching element 16.

The power conversion device 100 is controlled by the control device 104. In the present embodiment, the current state value of power conversion device 100 is input to control device 104, and control device 104 controls power conversion device 100 according to the input state value. The state value is controlled by the voltage v _b of the DC power source 10, the reactor current i _L flowing through the reactor 12, the capacitor voltage v _c across the capacitor 18, the motor currents i _u and i _{w as} loads, and the motor rotation angle θ. Input to the device 104. Controller 104 calculates the output current i _m of the power conversion device 100 from the current i _u, i rotation angle of _w and motor θ of the motor.

  FIG. 2 is a diagram illustrating the configuration of the control device 104. The control device 104 includes a PI controller 20, an inductance setting unit 22, a prediction time calculator 24, a model prediction controller (MPC) 26, and a limiter 28.

The PI controller 20 receives the deviation between the capacitor voltage command value v _c ^*, which is a command value for control of the power converter 100, and the capacitor voltage v _c , which is the current voltage value of the capacitor 18, and performs control of the MPC 26. The capacitor voltage command value v _c ^** as the target value is output. The capacitor voltage command value v _c ^** is input to the MPC 26.

The inductance setting unit 22 obtains the current value (inductance) of the reactor 12 based on the current reactor current i _L. In the power conversion apparatus 100, the inductance L of the reactor 12 depending on the reactor current i _L is changed. FIG. 3 is a diagram illustrating a change in the inductance L of the reactor 12 with respect to the current value. In FIG. 3, the current value on the horizontal axis is normalized with the maximum current as 1, and the inductance L on the vertical axis is normalized with 1 when the current value is 0. Inductance setter 22 sets the inductance L outputs corresponding to reactor current i _L based on the relational expression or database indicating the relationship between the reactor current i _L and the inductance L in FIG. The inductance L is input to the prediction time calculator 24 and the MPC 26.

The prediction time calculator 24 calculates a prediction time used when the state value of the power conversion device 100 is predicted in the MPC 26. That is, in the prediction time calculator 24 performs processing for calculating the time until the capacitor voltage v _c of the capacitor 18 at the time of changing the duty ratio d of the power conversion device 100 changes as the estimated time.

Relationship of Equation (1) and the capacitor voltage v _c and capacitor current i _c, equation a capacitor current i _c relation of equation (2) of the load current i _m, and reactor current i _L and the capacitor voltage v _c It is represented by the relational expression (3). Here, the value of capacitor 18 (capacity) C, the value of the reactor 12 (inductance) L, the estimated time T _p. As the inductance L, a value obtained by the inductance setting unit 22 is applied.

Note that k indicates the number of times of control. For example, d (k) represents the duty ratio d in the kth control, and d (k + 1) represents the duty ratio d in the (k + 1) th control. The same applies to other state quantities.

Differentiating equation (1) yields equation (4).

Substituting equation (2) into equation (4) yields equation (5).

Assuming that the load current i _m is not changed in Equation (5) becomes equation (6).

By substituting Equation (7) obtained by transforming Equation (3) into Equation (6), Equation (8) is obtained.

When formula (8) is organized and combined into a quadratic equation for the expected time T _p , formula (9) is obtained.

Here, since the expected time T _p can be expressed as a product of the control cycle T _c and the number of steps n (T _p = T _c × n), Equation (9) is a quadratic equation for the number of steps n. It can be expressed as Equation (10). In addition, the coefficient of each term in Formula (10) is set to a, b, c, respectively.

When the quadratic equation is solved for the number of steps n, Equation (11) is obtained. Since the number of steps n is always a positive number, only the + sign in the equation (11) may be calculated as the number of steps n. The expected time T _p is calculated by multiplying the number of steps n by the control period T _c . The expected time T _p is input to the MPC 26.

  In the MPC 26, using the state equation of the power conversion device 100, a predetermined state value (in the power conversion device 100 when the duty ratio d of the first switching element 14 and the second switching element 16 is changed to a plurality of different values ( A predicted value for the state quantity is calculated, and the power conversion apparatus 100 is controlled according to the difference between the command value indicating the target of the state value (state quantity) and the predicted value.

In the present embodiment, the MPC 26 calculates a capacitor voltage predicted value v _c ^{^} (hat) that is a predicted value of the voltage of the capacitor 18 as a predetermined state value. Then, the MPC 26 performs processing for obtaining a duty ratio d (= d (k + 1)) that becomes a capacitor voltage predicted value v _c ^{^} (hat) that approaches the capacitor voltage command value v _c ^** .

  As shown in FIG. 4, the MPC 26 includes an adder 30 (30-2 to 30-129), a prediction calculator 32 (32-1 to 32-129), and an evaluation function calculator 34 (34-1 to 34-129). ) And a minimum value selector 36.

  The adder 30 (30-2 to 30-129) changes the duty ratio d (k) by adding a predetermined value to the current duty ratio d (k) and outputs the change. In the present embodiment, it is assumed that the duty ratio d (k) is expressed in a range of values from 0 to 1023. That is, it is assumed that the duty ratio d is 0 when the second switching element 16 that is the lower arm is always on and the first switching element 14 that is the upper arm is always off. Further, it is assumed that the duty ratio d when the second switching element 16 that is the lower arm is always off and the first switching element 14 that is the upper arm is always on is represented by 1023. The adder 30 gives a change in the range of d (k) ± 64 with the current duty ratio d (k) as the center value, and outputs the change. The range of the change is preferably set based on the dead time period and the PWM cycle of the power conversion device 100. For example, it is preferable that the range of conversion is larger than the value calculated by the numerical range of dead time / PWM period × duty ratio d. Specifically, when the dead time is 5 μs and the PWM cycle is 100 μs, when the duty ratio d is expressed in the range of 0 to 1023, a numerical range larger than 5/100 × 1023 = 51 is set as the range of change. It is preferable to do. On the other hand, in order to make the calculation load as small as possible, it is preferable that the range of change is as narrow as possible. Therefore, in this embodiment, an example in which the range of change is ± 64 is shown.

  The adder 30-2 adds 1 to the current duty ratio d (k) and outputs d (k) +1. The adder 30-3 adds 2 to the current duty ratio d (k) and outputs d (k) +2. Similarly, adders 30-4 to 30-65 add 3 to 64 to the current duty ratio d (k) and output the result. The adder 30-66 subtracts 1 from the current duty ratio d (k) and outputs d (k) -1. The adder 30-67 subtracts 2 from the current duty ratio d (k) and outputs d (k) -2. Similarly, adders 30-68 to 30-129 subtract 3 to 64 from the current duty ratio d (k) and output the result. Outputs from the adders 30-2 to 30-129 are input to the prediction calculators 32-2 to 32-129, respectively.

Prediction calculator 32, the output from the adder 30, the capacitor voltage v _c, reactor current i _L, the power supply voltage v _b, the output current (load current) i _m and expected time T _p using capacitor voltage prediction value v _c ^{^} (Hat) is calculated and output.

Here, the state equation of the power conversion device 100 is expressed by Equation (12).

When the equation (12) is discretized using the bilinear transformation, the equation (13) is obtained.

Capacitor voltage predicted value v _c ^{^} (hat) is an expression obtained by replacing the first row v _c (k + 1) on the left side with v _c ^{^} [d (k) + a] (hat) and T replaced with predicted time T _p ( 14).

The prediction calculator 32-1 calculates and outputs v _c ^{^} [d (k)] (hat) with a in Equation (14) as 0. The prediction calculator 32-2 calculates and outputs v _c ^{^} [d (k) +1] (hat), where a in Equation (14) is 1. Similarly, the prediction calculator 32-3～ prediction calculator 32-65 is, v _c ^{^} as 2 to 64 of a respective [d (k) + a] calculates and outputs the (hat). The prediction computing unit 32-66 calculates and outputs v _c ^{^} [d (k) −1] (hat), where a in Equation (14) is −1. The prediction computing unit 32-67 calculates and outputs v _c ^{^} [d (k) −2] (hat), where “a” in Expression (14) is −2. Similarly, the prediction calculators 32-68 to 32-129 calculate and output v _c ^{^} [d (k) + a] (hat), where a is −3 to −64, respectively. The outputs of the prediction calculators 32-1 to 32-129 are input to the evaluation function calculators 34-1 to 34-129, respectively.

The evaluation function calculator 34 calculates the evaluation function J based on the capacitor voltage command value v _c ^** input from the PI controller 20 and the capacitor voltage predicted value v _c ^{^} (hat) input from the prediction calculator 32. And output the calculation result. The evaluation function J is expressed by Equation (15).

  The evaluation function calculator 34-1 calculates J [d (k)] by setting a in Equation (15) to 0 and outputs the result. The evaluation function calculator 34-2 calculates and outputs J [d (k) +1], where a in Equation (15) is 1. Similarly, the evaluation function calculator 34-3 to the evaluation function calculator 34-65 calculate and output J [d (k) + a] with a being 2 to 64, respectively. The evaluation function calculator 34-66 calculates and outputs J [d (k) −1], where a in Equation (15) is −1. The evaluation function calculator 34-67 calculates and outputs J [d (k) -2], where a in Equation (15) is −2. Similarly, the evaluation function calculators 34-68 to 34-129 calculate and output J [d (k) + a], where a is −3 to −64, respectively. The outputs of the evaluation function calculators 34-1 to 34-129 are input to the minimum value selector 36.

  The minimum value selector 36 includes J [d (k)], J [d (k) +1]... J [d () calculated by the evaluation function calculator 34-1 to evaluation function calculator 34-129. k) −64], the minimum value is selected, and d (k) + a having the evaluation function J as the minimum value is output as the duty ratio d (k + 1) in the next control.

  The limiter 28 receives the duty ratio d (k + 1) output from the MPC 26 and limits the input duty ratio d (k + 1) to be within the optimum duty ratio range DR.

FIG. 5 shows the relationship between the output P from the power converter 100 and the duty ratio d. As shown in FIG. 5, in order to set the output P to the maximum output Pmax, the power conversion apparatus 100 may be controlled so that the duty ratio d becomes the power supply voltage v _b / (2 × capacitor voltage v _c ). On the other hand, if the power conversion apparatus 100 is controlled such that the duty ratio d is lower than the power supply voltage v _b / (2 × capacitor voltage v _c ), the output P is lowered. Therefore, it is preferable to control the power conversion apparatus 100 so that the duty ratio d is in a range with the power supply voltage v _b / (2 × capacitor voltage v _c ) as a lower limit. The optimum duty ratio range DR does not necessarily need to be set to a range in which the power supply voltage v _b / (2 × capacitor voltage v _c ) is a lower limit, and is smaller than the maximum output Pmax, for example, as shown in FIG. The duty ratio d _L when the output value Pa is set as the output upper limit may be set as the lower limit. Furthermore, the duty ratio d _H when the output value Pb is set as the regeneration upper limit may be set as the upper limit of the optimum duty ratio range DR.

  If the duty ratio d (k + 1) from the MPC 26 is a value within the set optimum duty ratio range DR, the limiter 28 outputs the duty ratio d (k + 1) output from the MPC 26 as it is. Further, the limiter 28 determines that the duty ratio d (k + 1) output from the MPC 26 is within the optimum duty ratio range DR if the duty ratio d (k + 1) from the MPC 26 is outside the set optimum duty ratio range DR. Limit output to fit.

The control device 104 controls the ON period of the first switching element 14 and the second switching element 16 so that the duty ratio d output from the limiter 28 is obtained. Thus, the power conversion device 100, the capacitor voltage v _c so that the capacitor voltage command value v _c ^** which is a command value is controlled.

In the present embodiment, the control is performed based on the capacitor voltage command value v _c ^**, but the present invention is not limited to this. Or as an embodiment which performs control based on the reactor current command value i _L ^* in place of the capacitor voltage instruction value v _c ^**.

In this case, the PI controller 20 receives a deviation between the capacitor voltage command value v _c ^* which is a command value for control of the power converter 100 and the capacitor voltage v _c which is the current voltage value of the capacitor 18, and the MPC 26 The reactor current command value i _L ^*, which is the target value for the control, is output. Reactor current command value i _L ^* is input to MPC 26.

Further, the MPC 26 calculates a reactor current predicted value i _L ^{^} (hat) that is a predicted value of the current flowing through the reactor 12 as a predetermined state value. Then, the MPC 26 performs a process for obtaining a duty ratio d (= d (k + 1)) that becomes the reactor current predicted value i _L ^{^} (hat) that approaches the reactor current command value i _L ^* .

Specifically, by using the prediction calculator 32, the output from the adder 30, the capacitor voltage v _c, reactor current i _L, the power supply voltage v _b, the output current (load current) i _m and expected time T _p reactor The current predicted value i _L ^{^} (hat) is calculated and output. Reactor current predicted value i _L ^{^} (hat) is replaced by i _L ^{^} [d (k) + a] (hat) in the second line i _L (k + 1) on the left side of equation (13), and T is predicted time T _p . reactor current predicted value using equation (16) obtained by developing Te ^{_{i L ^ [d (k)}} + a] may be calculated (hat).

The evaluation function calculator 34 calculates the evaluation function J based on the reactor current command value i _L ^* input from the PI controller 20 and the reactor current predicted value i _L ^{^} (hat) input from the prediction calculator 32. And output the calculation result. The evaluation function J is expressed by Equation (17).

  The minimum value selector 36 includes J [d (k)], J [d (k) +1]... J [d () calculated by the evaluation function calculator 34-1 to evaluation function calculator 34-129. k) −64], the minimum value is selected, and d (k) + a having the evaluation function J as the minimum value is output as the duty ratio d (k + 1) in the next control. The limiter 28 receives the duty ratio d (k + 1) output from the MPC 26 and limits the input duty ratio d (k + 1) to be within the optimum duty ratio range DR. The control device 104 controls the power conversion device 100 based on the duty ratio d (k + 1) obtained in this way.

In the present embodiment, the value obtained by multiplying the number of steps n obtained as a solution of Equation (11) by the control period T _c is defined as the predicted time T _p , but the predicted time T _p is determined by adding the number of steps n to the control period. It is sufficient that the value is equal to or greater than a value obtained by multiplying _Tc .

FIGS. 7A to 7C show the capacitor voltage command value v _c ^** , the capacitor voltage v _c, and the power supply voltage when the power conversion device 100 is controlled by the control device 104 in the first embodiment, respectively. v _b is shown. FIG. 8 shows the reactor current i _L when the power conversion device 100 is controlled by the control device 104 in the first embodiment. Moreover, FIG. 9 shows the solution of Formula (11).

  In FIGS. 7, 8 and 9, the time on the horizontal axis indicates the same time range. In FIGS. 7A to 7C, the voltage values on the vertical axis indicate the same voltage range.

By the control device 104 of the present embodiment, as shown in FIGS. 7-9, the power conversion apparatus 100 can be controlled so that the capacitor voltage instruction value v _c ^** the capacitor voltage v _c is the command value to follow .

On the other hand, FIG. 10A to FIG. 10C show capacitor voltage command values v _c ^** when the power conversion device 100 is controlled using a time shorter than the expected time T _p in the first embodiment. shows the capacitor voltage v _c and the power supply voltage v _b. FIG. 11 shows reactor current i _L when power converter 100 is controlled using a time shorter than expected time T _p in the first embodiment. FIG. 12 shows a solution of Equation (11) when a time shorter than the expected time T _p in the first embodiment is used.

  In FIGS. 10, 11, and 12, the time on the horizontal axis indicates the same time range. 10A to 10C, the voltage values on the vertical axis indicate the same voltage range.

When using a shorter time than the expected time T _p in the first embodiment, as shown in FIGS. 7-9, the power conversion apparatus 100 includes the capacitor voltage instruction value v of the capacitor voltage v _c is the command value _c ^** is not sufficiently followed, and the stability of the control is reduced as compared with the first embodiment.

<Second Embodiment>
In the first embodiment, the configuration of the control device 104 that measures the reactor current i _L that actually flows through the reactor 12 and controls the duty ratio d based on the reactor current i _L has been described. In the second embodiment, the power conversion device 200, as shown in FIG. 13, not provided with a sensor for measuring the reactor current i _L flowing through the reactor 12. Alternatively, as shown in FIG. 14, the control in the device observer 40 contained in the 104a, is an estimate of the current flowing through the error duty ratio estimate ^{Δd ~ (= Δd ~ (k} ) ( tilde)) and the reactor 12 Reactor current estimated value i _L ^~ (tilde) is calculated and output.

The state equation of the power conversion device 200 is expressed by Equation (12). When the error duty ratio Δd considering the dead time is incorporated in Equation (12), the state equation shown in Equation (18) is obtained.

When the equation (18) is discretized using the bilinear transformation, the equation (19) is obtained.

Based on Equation (19), a capacitor voltage predicted value (tilde) v _c ^˜ (k) that is a predicted value of the voltage of the capacitor 18, a reactor current predicted value (tilde) i _L that is a predicted value of the current of the reactor 12 ^˜ (K) and the error duty ratio predicted value (tilde) Δd ^to (k), which is a predicted value of the error duty ratio, can be expressed as Equation (20). Here, it is the control cycle T, and h _{1 to} h ₃ are proportional constants.

Observer 40 is input capacitor voltage command value v _c ^*, capacitor voltage v _c, reactor current i _L, by substituting the power supply voltage v _b and the output current i _m in equation (20), current error duty ratio An estimated value of Δd (= Δd (k)) is calculated. The estimated error duty ratio estimated values Δd ^to (= Δd ^to (k) (tilde)) can be calculated by replacing k in the equation (20) with k−1. The calculated error duty ratio estimate ^{Δd ~ (= Δd ~ (k} ) ( tilde)) is input to MPC26a.

Moreover, the observer 40, the capacitor voltage instruction value v _c ^*, capacitor voltage v _c, the power supply voltage v _b and by substituting the output current i _m the formula (20), reactor current estimated value i _L ^~ (= i _L to calculate ^- the (k) (tilde)). Reactor current estimated values i _L ^to (= i _L ^to (k) (tilde)) can be calculated by replacing k in equation (20) with k−1 and processing. The calculated reactor current estimated value i _L ^˜ (= i _L ^˜ (k) (tilde)) is input to the inductance setting device 22a and the MPC 26a.

Instead of using the current reactor current i _L in the inductance setter 22, the inductance setter 22 a is based on the reactor current estimated value i _L ^˜ (= i _L ^˜ (k) (tilde)). A value of 12 (inductance) is obtained. The inductance setting device 22a corresponds to the reactor current estimated value i _L ^˜ (= i _L ^˜ (k) (tilde)) based on the relational expression or database indicating the relation between the reactor current i _L and the inductance L in FIG. Set inductance L and output. The inductance L is input to the prediction time calculator 24 and the MPC 26a.

The prediction time calculator 24 calculates a prediction time T _p used when the state value of the power conversion device 200 is predicted in the MPC 26a. Since the processing in the prediction time calculator 24 is the same as that in the first embodiment, description thereof is omitted.

  In the MPC 26a, using the state equation of the power conversion device 200, a predetermined state value (in the power conversion device 100 when the duty ratio d of the first switching element 14 and the second switching element 16 is changed to a plurality of different values ( The predicted value for the state quantity) is calculated, and the power conversion device 200 is controlled according to the difference between the command value indicating the target of the state value (state quantity) and the predicted value.

In the present embodiment, the MPC 26a calculates a capacitor voltage predicted value v _c ^{^} (hat) that is a predicted value of the voltage of the capacitor 18 as a predetermined state value. Then, the MPC 26a performs processing for obtaining a duty ratio d (= d (k + 1)) that becomes a capacitor voltage predicted value v _c ^{^} (hat) that approaches the capacitor voltage command value v _c ^** .

  As shown in FIG. 15, the MPC 26a includes an adder 30 (30-2 to 30-129), a prediction calculator 32a (32a-1 to 32a-129), and an evaluation function calculator 34a (34a-1 to 34a-129). ) And a minimum value selector 36.

  The adder 30 (30-2 to 30-129) performs the same processing as that of the first embodiment, and thus description thereof is omitted.

The predictive calculator 32a outputs the output from the adder 30, the capacitor voltage v _c , the reactor current estimated value i _L ^˜ (= i _L ^˜ (k) (tilde)), the power supply voltage v _b , the output current (load current) i. _The capacitor voltage predicted value v _c ^{^} (hat) is calculated and output using _m , the error duty ratio estimated value Δd ^˜ (= Δd ^˜ (k) (tilde)) and the expected time T _p . The predictive calculator 32a replaces the reactor current i _L with the reactor current estimated value i _L ^˜ (= i _L ^˜ (k) (tilde)) and the error duty ratio estimated value Δd ^˜ (= Δd ^˜ (k) (tilde). )) Is different from the predictive calculator 32 of the first embodiment.

In the formula (19), the capacitor voltage predicted value v _c ^{^} (hat) is obtained by changing the first line v _c (k + 1) on the left side to v _c ^{^} [d (k) + a] (hat) and i _L (k) on the right side. Is calculated using Equation (21) in which i _L ^˜ (= i _L ^˜ (k) (tilde)) and T is replaced with the predicted time T _p .

The predictive calculator 32a-1 calculates and outputs v _c ^{^} [d (k)] (hat) with a in Equation (21) as 0. The prediction computing unit 32a-2 calculates and outputs v _c ^{^} [d (k) +1] (hat), where a in Equation (21) is 1. Similarly, the prediction calculator 32a-3 to the prediction calculator 32a-65 calculate and output v _c ^{^} [d (k) + a] (hat), where a is 2 to 64, respectively. The prediction calculator 32a-66 calculates and outputs v _c ^{^} [d (k) −1] (hat), where “a” in Expression (21) is −1. The prediction calculator 32a-67 calculates and outputs v _c ^{^} [d (k) −2] (hat), where “a” in Expression (21) is −2. Similarly, the prediction calculators 32a-68 to 32a-129 calculate and output v _c ^{^} [d (k) + a] (hat), where a is −3 to −64, respectively. The outputs of the prediction calculators 32a-1 to 32a-129 are input to the evaluation function calculators 34a-1 to 34a-129, respectively.

The evaluation function calculator 34a calculates the evaluation function J based on the capacitor voltage command value v _c ^** input from the PI controller 20 and the capacitor voltage predicted value v _c ^{^} (hat) input from the prediction calculator 32a. And output the calculation result. The evaluation function J may use the formula (15).

  The evaluation function calculator 34a-1 calculates J [d (k)] by setting a in Equation (15) to 0 and outputs it. The evaluation function calculator 34a-2 calculates and outputs J [d (k) +1], where a in Equation (15) is 1. Similarly, the evaluation function calculator 34a-3 to the evaluation function calculator 34a-65 calculate and output J [d (k) + a], where a is 2 to 64, respectively. The evaluation function calculator 34a-66 calculates and outputs J [d (k) −1], where a in Formula (15) is −1. The evaluation function calculator 34a-67 calculates and outputs J [d (k) -2], where a in Equation (15) is -2. Similarly, the evaluation function calculators 34a-68 to 34a-129 calculate and output J [d (k) + a], where a is −3 to −64, respectively. The outputs of the evaluation function calculators 34 a-1 to 34 a-129 are input to the minimum value selector 36.

In the present embodiment, it is more preferable to use Equation (22) as the evaluation function J. In this case as well, J [d (k)], J [d (k) +1]... J [d (k) -64] are used in the evaluation function calculators 34a-1 to 34a-129, respectively. Is calculated and output.

  The minimum value selector 36 includes J [d (k)], J [d (k) +1]... J [d () calculated by the evaluation function calculators 34a-1 to 34a-129. k) −64], the minimum value is selected, and d (k) + a having the evaluation function J as the minimum value is output as the duty ratio d (k + 1) in the next control.

  The limiter 28 receives the duty ratio d (k + 1) output from the MPC 26 and limits the input duty ratio d (k + 1) to be within the optimum duty ratio range DR. The control device 104a controls the power conversion device 200 based on the duty ratio d (k + 1) obtained in this way.

Note that the observer 40 may further calculate capacitor voltage estimated values v _c ^to (v _c ^to (k)) (tilde) that are estimated values of the voltage of the capacitor 18. In this case, the prediction calculator 32a calculates and outputs a capacitor voltage predicted value v _c ^{^} (hat) using the capacitor voltage estimated values v _c ^to (v _c ^to (k)) (tilde). Capacitor voltage predicted value v _c ^{^} (hat) is _{calculated by} ^changing v _c ^{^} [d (k) + a] (hat) in the first row v _c (k + 1) on the left side and i _L (k) on the right side in equation (19). the ^{_{i L ~ (= i L ~}} (k) ( tilde)), v _c a ^{_{(k) v c ~ (=}} v c ~ (k) ( tilde)), was developed replaced by the time predicting the T T _p Calculated using Equation (23).

In the present embodiment, the control is performed based on the capacitor voltage command value v _c ^**, but the present invention is not limited to this. Or as an embodiment which performs control based on the reactor current command value i _L ^* in place of the capacitor voltage instruction value v _c ^**.

In this case, the PI controller 20 receives the deviation between the capacitor voltage command value v _c ^*, which is a command value for control of the power converter 200, and the capacitor voltage v _c , which is the current voltage value of the capacitor 18, and the MPC 26a The reactor current command value i _L ^*, which is the target value for the control, is output. Reactor current command value i _L ^* is input to MPC 26a.

Further, the MPC 26a calculates a reactor current predicted value i _L ^{^} (hat) that is a predicted value of the current flowing through the reactor 12 as a predetermined state value. Then, the MPC 26a performs a process of obtaining a duty ratio d (= d (k + 1)) that becomes a reactor current predicted value i _L ^{^} (hat) that approaches the reactor current command value i _L ^* .

Specifically, in the predictive calculator 32a, the output from the adder 30, the capacitor voltage v _c , the reactor current estimated value i _L ^˜ (= i _L ^˜ (k) (tilde)), the power supply voltage v _b , the output current Reactor current predicted value i _L ^{^} (hat) is calculated and output using (load current) i _m , error duty ratio estimated value Δd ^~ (= Δd ^~ (k) (tilde)) and expected time T _p . Reactor current predicted value i _L ^{^} (hat) is the equation second line of the left-hand side of (19) i _L a ^{_{(k + 1) i L ^}} [d (k) + a] ( hat), the right-hand side of the i _L (k) i _L ^˜ (= i _L ^˜ (k) (tilde)), the reactor current prediction value i _L ^{^} [d (k) + a] (hat) is calculated using a mathematical expression expanded by replacing T with the prediction time T _p do it.

The evaluation function calculator 34 calculates the evaluation function J based on the reactor current command value i _L ^* input from the PI controller 20 and the reactor current predicted value i _L ^{^} (hat) input from the prediction calculator 32a. And output the calculation result. The evaluation function J may use the formula (17).

Note that it is more preferable to use Equation (24) as the evaluation function J. In this case as well, J [d (k)], J [d (k) +1]... J [d (k) -64] are used in the evaluation function calculators 34a-1 to 34a-129, respectively. Is calculated and output.

  The minimum value selector 36 includes J [d (k)], J [d (k) +1]... J [d () calculated by the evaluation function calculator 34-1 to evaluation function calculator 34-129. k) −64], the minimum value is selected, and d (k) + a having the evaluation function J as the minimum value is output as the duty ratio d (k + 1) in the next control. The limiter 28 receives the duty ratio d (k + 1) output from the MPC 26 and limits the input duty ratio d (k + 1) to be within the optimum duty ratio range DR. The control device 104 controls the power conversion device 100 based on the duty ratio d (k + 1) obtained in this way.

Further, in the present embodiment, the control is performed using the error duty ratio estimated value Δd ^to (= Δd ^to (k) (tilde)) and the reactor current estimated value i _L ^to (tilde). If a sensor for measuring the reactor current i _L is provided as shown in FIG. 1, the reactor current i _L may be used without estimating the reactor current estimated value i _L ^˜ (tilde).

In this case, the prediction calculator 32a replaces the estimated reactor current i _L ^˜ (tilde) on the right side of the equation (21) with the actually measured reactor current i _L to replace the estimated capacitor voltage v _c ^{^} (hat). What is necessary is just to calculate. Further, in the predictive calculator 32a, the reactor current estimated value i _L ^{^} (hat) is calculated by replacing the reactor current estimated value i _L ^˜ (tilde) on the right side of Equation (23) with the actually measured reactor current i _L. do it.

  Also in power conversion device 200 in the present embodiment, stable control similar to that in the first embodiment can be realized by applying control by control device 104a.

<Third Embodiment>
In the third embodiment, the configuration of the MPC 26a is changed. In the present embodiment, the MPC 26a transforms the state equation of the power converter 200 into a quadratic equation for the duty ratio d of the first switching element 14 and the second switching element 16, and is calculated by the observer 40 as the quadratic equation. The duty ratio d is calculated by applying the estimated reactor current value i _L ^~ (tilde), the capacitor voltage estimated value v _c ^~ (tilde), and the error duty ratio estimated value Δd ^~ (tilde). The control device 104a controls the power conversion device 200 using the calculated duty ratio d.

The second line i _L (k + 1) on the left side of Equation (19) is i _L ^{^ *} (k), d (k) on the right side is Δd ^to (k) (tilde), and the reactor current i _L is the reactor current estimated value i. _L ^to (k) (tilde), capacitor voltage v _c is replaced with capacitor voltage estimated value v _c ^to (k) (tilde), T is replaced with prediction time T _p and changed to a quadratic equation for duty ratio d (k) Then, Equation (25) is obtained.

When the quadratic equation of Equation (25) is solved with respect to the duty ratio d (k + 1), it is expressed by Equation (26).

The MPC 26a uses the equation (26) to calculate each state quantity of the power converter 200 and the reactor current estimated value i _L ^~ (tilde) calculated by the observer 40, the capacitor voltage estimated value v _c ^~ (tilde), and the error duty ratio estimation. The duty ratio d (k + 1) used for control is calculated by substituting the values Δd ^to (tilde).

The limiter 28 receives the duty ratio d (k + 1) output from the MPC 50, and limits the input duty ratio d (k + 1) to be within the optimum duty ratio range DR. The control device 104a controls the ON period of the first switching element 14 and the second switching element 16 so that the duty ratio d (= d (k + 1)) output from the limiter 28 is obtained. Thereby, the power converter 200 controls the capacitor voltage v _c and the reactor current i _L so that the capacitor voltage command value v _c ^* and the reactor current command value i _L ^*, which are command values, are obtained.

Also, the duty ratio d (k + 1) can be obtained by another method. The first line v _c (k + 1) on the left side of Equation (19) is v _c ^{^ *} (k), d (k) on the right side is Δd ^to (k) (tilde), and the reactor current i _L is the estimated reactor current i. _L ^to (k) (tilde), capacitor voltage v _c is replaced with capacitor voltage estimated value v _c ^to (k) (tilde), T is replaced with prediction time T _p and changed to a quadratic equation for duty ratio d (k) Then, Equation (27) is obtained.

When the quadratic equation of Equation (27) is solved for the duty ratio d (k + 1), it is expressed by Equation (28).

The MPC 26a uses Equation (28) to calculate each state quantity of the power converter 200 and the reactor current estimated value i _L ^~ (tilde), the capacitor voltage estimated value v _c ^~ (tilde), and the error duty ratio estimation calculated by the observer 40. The duty ratio d (k + 1) used for control is calculated by substituting the values Δd ^to (tilde). The control device 104a controls the power conversion device 200 using the calculated duty ratio d (k + 1).

Further, the duty ratio d (k + 1) can be obtained by another method. In MPC26a, instead of the reactor current estimated value i _L ^~ (tilde) and capacitor voltage estimated value v _c ^~ (tilde), may be respectively applied to the measured value of the reactor current i _L and capacitor voltage v _c. In this case, it may be provided sensors for measuring the reactor current i _L and capacitor voltage v _c. In this case, Equation (19) is changed to a quadratic equation as shown in Equation (29).

When the quadratic equation of Equation (29) is solved with respect to the duty ratio d (k + 1), it is expressed by Equation (30).

MPC26a is in equation (28), the duty ratio used for control by substituting each state quantity and the reactor current i _L of the power conversion device 200, the capacitor voltage v _c and the error duty ratio estimate [Delta] d ^~ a (tilde) d ( k + 1) is calculated. The control device 104a controls the power conversion device 200 using the calculated duty ratio d (k + 1).

In these methods, the power inverter 200, the capacitor voltage command is a command value value v _c ^* and reactor current command value i _L ^* become as capacitor voltage v _c and reactor current i _L is controlled.

<Fourth embodiment>
In the first to third embodiments, the aspect in which the prediction time T _p for predicting the state value in the MPC 26 or MPC 26a is optimized has been described. In contrast, in the fourth embodiment, a mode in which the settling time T _s for setting the command value for the state value is optimized will be described.

  FIG. 16 shows a basic configuration of power conversion device 300 according to the embodiment of the present invention. The power conversion device 300 includes a DC power supply 10, a reactor 12, a first switching element 14, a second switching element 16, and a capacitor 18. Since the configurations of the DC power supply 10, the reactor 12, the first switching element 14, the second switching element 16, and the capacitor 18 are the same as those in the first to third embodiments, the description thereof is omitted.

The power conversion device 300 is controlled by the control device 104b. In the present embodiment, the current state value of power conversion device 300 is input to control device 104b, and control device 104b controls power conversion device 300 according to the input state value. The state value is controlled by the voltage v _b of the DC power source 10, the reactor current i _L flowing through the reactor 12, the capacitor voltage v _c across the capacitor 18, the motor currents i _u and i _{w as} loads, and the motor rotation angle θ. Input to the device 104. Controller 104 calculates the output current i _m of the power conversion device 100 from the current i _u, i rotation angle of _w and motor θ of the motor.

  FIG. 17 is a diagram illustrating a configuration of the control device 104b. The control device 104b includes an observer 40, a command value generator 50, an inductance setting unit 22a, a settling time calculator 52, a model prediction controller (MPC) 26b, and a limiter 28.

Observer 40, as in the second embodiment, the input capacitor voltage command value v _c ^*, capacitor voltage v _c, reactor current i _L, the power supply voltage v _b and the output current i _m in equation (20) by substituting for, calculates an error duty ratio estimate ^{Δd ~ (= Δd ~ (k} ) ( tilde)) and the reactor current estimated value ^{_{i L ~ (= i L ~}} (k) ( tilde)). The processing for calculating the error duty ratio estimated value Δd ^to (= Δd ^to (k) (tilde)) and the reactor current estimated value i _L ^to (= i _L ^to (k) (tilde)) is the second embodiment. Since it is the same as that of FIG.

Inductance setter 22a determines the reactor current estimated value i _L ^~ calculated in observer 40 the current value of the reactor 12 based on ^(tilde) (inductance). Since this process is the same as that of the second embodiment, a description thereof will be omitted. The inductance L is input to the settling time calculator 52 and the MPC 26b.

The settling time calculator 52 calculates a settling time used in the command value generator 50 to set a command value for the state quantity. That is, the settling time calculator 52 performs a process of calculating, as the settling time, the time until the reactor current command value i _L ^* that becomes the command value when performing control to change the duty ratio d of the power converter 300. Do. In the present embodiment, a settling time T _{s serving as} a reference for setting the reactor current command value i _L ^* will be described.

By substituting Equation (31) obtained by transforming Equation (3) into Equation (6), Equation (32) is obtained. The predicted time T _p is replaced with the settling time T _s .

When formula (32) is rearranged and rewritten into a quadratic equation for settling time T _s , formula (33) is obtained.

Here, since the settling time T _s can be expressed as a product of the control cycle T _c and the number of steps n (T _s = T _c × n), Equation (33) is a quadratic equation for the number of steps n. It can be expressed as Equation (34). Note that the coefficients of the terms in the equation (34) are a, b, and c, respectively.

When the quadratic equation is solved for the number of steps n, Equation (35) is obtained. Since the number of steps n is always a positive number, only the + sign in the equation (35) may be calculated as the number of steps n. The settling time T _s is calculated by multiplying the number of steps n by the control period T _c . The settling time T _s is input to the command value generator 50 and the MPC 26b.

In command value generator 50, reactor current command value i _L ^* is set based on settling time T _s . In Equation (19), v _c (k + 1) in the first row on the left side is v _c ^* (k), and i _L (k) in the second term, second row and first column on the right side is i _L ^* (k). When Δd (k) in the second term, third row, first column is replaced with Δd ^to (k) (tilde), and T is replaced with settling time T _s , the first row is rearranged to obtain Equation (36). Command value generator 50 sets reactor current command value i _L ^* (k) based on Equation (36), and outputs it to MPC 26b. In this embodiment, applying the reactor current estimated value i _L ^~ instead of the reactor current i _L ^(k) (tilde).

In the MPC 26b, using the state equation of the power converter 300, a predetermined state value (in the power converter 300 when the duty ratio d of the first switching element 14 and the second switching element 16 is changed to a plurality of different values ( A predicted value for the state quantity) is calculated, and the power conversion apparatus 300 is controlled according to the difference between the command value of the state value (state quantity) set based on the settling time T _s and the predicted value.

In the present embodiment, the MPC 26b calculates a reactor current predicted value i _L ^{^} (hat) that is a predicted value of the reactor 12 as a predetermined state value. Then, the MPC 26b performs a process for obtaining a duty ratio d (= d (k + 1)) that becomes the reactor current predicted value i _L ^{^} (hat) that approaches the reactor current command value i _L ^* .

  As shown in FIG. 18, the MPC 26b includes an adder 30 (30-2 to 30-129), a prediction calculator 32b (32b-1 to 32b-129), and an evaluation function calculator 34b (34b-1 to 34b-129). ) And a minimum value selector 36.

  The adder 30 (30-2 to 30-129) performs the same processing as that of the first embodiment, and thus description thereof is omitted.

The predictive calculator 32b outputs the output from the adder 30, the reactor current command value i _L ^* from the command value generator 50, the capacitor voltage v _c , the reactor current estimated value i _L ^˜ (= i _L ^˜ (k) (tilde )), calculates the power supply voltage v _b, the output current (load current) i _m and the error duty ratio estimate ^{Δd ~ (= Δd ~ (k} ) ( reactor current predicted value using the tilde)) i _L ^{^} (hat) And output. The predictive calculator 32b replaces the reactor current i _L with the reactor current estimated value i _L ^˜ (= i _L ^˜ (k) (tilde)) and the error duty ratio estimated value Δd ^˜ (= Δd ^˜ (k) (tilde). )) And is different from the prediction calculator 32 in that processing is performed based on the reactor current command value i _L ^* from the command value generator 50.

In Equation (19), the reactor current prediction value i _L ^{^} (hat) is obtained by replacing i _L (k + 1) in the second row on the left side with i _L ^{^} [d (k) + a] (hat) and i _L (k) on the right side. Is replaced by i _L ^˜ (= i _L ^˜ (k) (tilde)), and the error duty ratio Δd is replaced with the error duty ratio estimated value Δd ^˜ (= Δd ^˜ (k) (tilde)). Is used to calculate.

The predictive calculator 32b-1 calculates and outputs i _L ^{^} [d (k)] (hat) by setting a in Equation (37) to 0. The prediction computing unit 32b-2 calculates and outputs i _L ^{^} [d (k) +1] (hat) with a in Equation (37) as 1. Similarly, the prediction calculator 32b-3 to the prediction calculator 32b-65 calculate and output i _L ^{^} [d (k) + a] (hat) with a being 2 to 64, respectively. The prediction calculator 32b-66 calculates and outputs i _L ^{^} [d (k) −1] (hat), where “a” in Expression (37) is −1. The prediction computing unit 32b-67 calculates and outputs i _L ^{^} [d (k) −2] (hat), where “a” in Expression (37) is −2. Similarly, the prediction calculators 32b-68 to 32b-129 calculate and output i _L ^{^} [d (k) + a] (hat), where a is −3 to −64, respectively. The outputs of the prediction calculators 32b-1 to 32b-129 are input to the evaluation function calculators 34b-1 to 34b-129, respectively.

The evaluation function calculator 34b calculates the evaluation function J based on the reactor current command value i _L ^* input from the command value generator 50 and the reactor current predicted value i _L ^{^} (hat) input from the prediction calculator 32b. And output the calculation result. The evaluation function J may use the formula (24). Also in the present embodiment, the mathematical expression (17) may be used as the evaluation function J.

  The evaluation function calculator 34b-1 calculates J [d (k)] by setting a in Equation (24) to 0 and outputs it. The evaluation function calculator 34b-2 calculates and outputs J [d (k) +1], where a in Equation (24) is 1. Similarly, the evaluation function calculator 34b-3 to the evaluation function calculator 34b-65 calculate and output J [d (k) + a] with a being 2 to 64, respectively. The evaluation function calculator 34b-66 calculates and outputs J [d (k) −1], where a in Formula (24) is −1. The evaluation function calculator 34b-67 calculates and outputs J [d (k) -2], where a in Equation (24) is -2. Similarly, the evaluation function calculators 34b-68 to 34b-129 calculate and output J [d (k) + a], where a is −3 to −64, respectively. The outputs of the evaluation function calculators 34b-1 to 34b-129 are input to the minimum value selector 36.

  The minimum value selector 36 includes J [d (k)], J [d (k) +1]... J [d () calculated by the evaluation function calculators 34b-1 to 34b-129. k) −64], the minimum value is selected, and d (k) + a having the evaluation function J as the minimum value is output as the duty ratio d (k + 1) in the next control.

The limiter 28 receives the duty ratio d (k + 1) output from the MPC 26b, and limits the input duty ratio d (k + 1) to be within the optimum duty ratio range DR. The control device 104b controls the ON period of the first switching element 14 and the second switching element 16 so that the duty ratio d (k + 1) output from the limiter 28 is obtained. Thus, the power conversion apparatus 100, reactor current i _L is controlled such that the reactor current command value i _L ^*.

19 (a) to 19 (c) show a capacitor voltage command value v _c ^* , a capacitor voltage v _c and a power supply voltage v when the power conversion device 300 is controlled by the control device 104b in the fourth embodiment. _b . 20A to 20B show the reactor current command value i _L ^* and the reactor current i _L when the power conversion device 300 is controlled by the control device 104b in the fourth embodiment. FIG. 21 shows a solution of Expression (35).

  In FIGS. 19, 20, and 21, the time on the horizontal axis indicates the same time range. In FIGS. 19A to 19C, the voltage values on the vertical axis indicate the same voltage range. 20A to 20B, the current values on the vertical axis indicate the same current range.

By the control device 104b of the present embodiment, as shown in FIGS. 19 to 21, the capacitor voltage instruction value the power converter 300 is a command value v _c ^* and reactor current command value i _L ^* to the capacitor voltage v _c And the reactor current i _L can be controlled to follow.

On the other hand, FIG. 22A to FIG. 22C show capacitor voltage command values v _c ^* when the power converter 300 is controlled using a time shorter than the settling time T _s in the fourth embodiment. It shows the capacitor voltage v _c and the power supply voltage v _b. FIG. 23A to FIG. 23B show the reactor current command value i _L ^* and the reactor current when the power conversion device 300 is controlled using a time shorter than the settling time T _s in the fourth embodiment. i _L is shown. FIG. 24 shows a solution of Expression (35) when a time shorter than the settling time T _S in the fourth embodiment is used.

  In FIGS. 22, 23, and 24, the time on the horizontal axis indicates the same time range. In FIGS. 22A to 22C, the voltage values on the vertical axis indicate the same voltage range. In FIGS. 23A to 23B, the current values on the vertical axis indicate the same current range.

When using a time shorter than the settling time T _S in the fourth embodiment, as shown in FIGS. 22 to 24, the power conversion device 300 uses the capacitor voltage command value v _c ^* and the reactor current command value i _L ^*. capacitor voltage v _c and reactor current i _L is not sufficiently follow, the stability of the control is decreased as compared with the fourth embodiment to.

Incidentally, when the sensor for measuring the reactor current i _L flowing through the reactor 12 actually provided, actually measured reactor instead of the reactor current estimated value i _L ^~ In the above-described embodiment ^(k) (tilde) Processing may be performed by applying the current i _L to each mathematical expression.

<Fifth embodiment>
Reactor current command value i _L ^* set by command value generator 50 can also be applied to the processing in the third embodiment. That is, the MPC 26a applies the reactor current command value i _L ^* set by the command value generator 50 and calculates the duty ratio when calculating the duty ratio d (k + 1) based on the quadratic equation of the formula (25). d (k + 1) is obtained. Thereafter, control is performed in the same manner as in the third embodiment based on the calculated duty ratio d (k + 1). The same applies to the case where control is performed based on Equation (27) and Equation (29).

  10 DC power supply, 12 reactor, 14 first switching element, 16 second switching element, 18 capacitor, 20 controller, 22, 22a inductance setting unit, 24 prediction time calculator, 26, 26a, 26b MPC, 28 limiter, 30 Adder, 32, 32a, 32b Prediction calculator, 34, 34a, 34b Evaluation function calculator, 36 Minimum value selector, 40 Observer, 50 Command value generator, 52 Settling time calculator, 100, 200, 300 Power conversion Device, 102 load, 104, 104a, 104b Control device.

Claims (6)

A control device for controlling a DC / DC converter,
A model prediction controller that predicts a predetermined state value in the prediction time in the DC / DC converter using the state equation of the DC / DC converter and controls the DC / DC converter according to the predicted state value When,
A prediction time calculator for setting the prediction time in the model prediction controller;
A control device comprising: The control device according to claim 1,
The said prediction time calculator sets the said prediction time using the value of the reactor and capacitor | condenser value which are contained in the said DC / DC converter, The control apparatus characterized by the above-mentioned. The control device according to claim 1,
The DC / DC converter includes a reactor connected to a power source, a first switching element and a second switching element that control a current flowing through the reactor, and a capacitor that smoothes an output voltage from the reactor,
The model predictive controller, the power supply voltage v _b of the power source, said predetermined condition value by applying the output current i _m of the capacitor voltage v _c and the DC / DC converter at both ends of the capacitor to the state equation A control apparatus that predicts and controls the DC / DC converter according to the predicted state value. The control device according to claim 3,
An observer for estimating an error duty ratio Δd with respect to a duty ratio d, which is a ratio of an ON period of the first switching element and the second switching element, according to a current state value of the DC / DC converter;
The model predictive controller, wherein the power supply voltage v _b, the capacitor voltage v _c, in addition to the output current i _m, predicts the predetermined state value by applying said error duty ratio Δd in the state equation A control device characterized by that. The control device according to claim 3,
According to the current state value of the DC / DC converter, an error duty ratio Δd with respect to a duty ratio d, which is a ratio of an ON period of the first switching element and the second switching element, and an estimated reactor current i flowing through the reactor with an observer to estimate the _L ^~,
The model predictive controller, wherein the power supply voltage v _b, the capacitor voltage v _c, in addition to the output current i _m, applying the error duty ratio Δd and the reactor current estimated value i _L ^~ in the state equation The control device predicts the predetermined state value by the following. It is a control device given in any 1 paragraph of Claims 1-4,
A control device configured to set a value of the reactor according to a current flowing through the reactor, and to apply the value of the reactor to the state equation.