Electric vehicle control method and electric vehicle control device

Abstract

This electric vehicle control method controls the operation of a rotary electric machine using a first electronic control unit and a second electronic control unit that communicates with the first electronic control unit via a communication network. The first electronic control unit calculates a basic command value for determining an operation state of the rotary electric machine, quantizes the basic command value with a quantization width corresponding to communication-load constraints in the communication network, and transmits a quantized basic command value, which is the basic command value that has been quantized, to the second electronic control unit via the communication network. The second electronic control unit receives the quantized basic command value from the first electronic control unit, calculates a correction command value for suppressing the vibration of an electric vehicle by using the quantized basic command value, and controls the operation of the rotary electric machine on the basis of the correction command value. When the basic command value is quantized in the first electronic control unit, quantization noise is shaped so that a predetermined frequency component including the resonance frequency of the vibration of the electric vehicle moves to the high frequency side.

Description

Control method for electric vehicle and control device for electric vehicle

[0001] The present invention relates to a control method and a control device for an electric vehicle that performs vibration damping control.

[0002] JP2014-128088A discloses an electric vehicle that prevents the promotion of torsional vibrations occurring in the drivetrain by processing the motor rotation speed with a low-pass filter when the limit value of the upper torque limiter is set using the motor rotation speed.

[0003] Conventionally, in electric vehicles, rotating electric machines are controlled to suppress vibrations occurring in the electric vehicle, such as torsional vibrations of the drive shaft. That is, in electric vehicles, vibration suppression control is performed when controlling the rotating electric machine. Specifically, the electric vehicle calculates a basic command value that determines the operating state of the rotating electric machine, for example, in accordance with the amount of accelerator operation by the driver. The electric vehicle does not control the rotating electric machine using this basic command value as is, but rather controls the rotating electric machine using a corrected command value that is corrected in accordance with the characteristics of the electric vehicle (vehicle characteristics such as torque transmission characteristics) so as to suppress vibrations. In this way, the rotating electric machine is driven while suppressing vibrations occurring in the electric vehicle.

[0004] However, even if the corrected command value is used as described above, vibrations may still occur in the electric vehicle. Specifically, when the control device for the electric vehicle is configured by multiple controllers interconnected via a communication bus and command values ​​used to control the rotating electric machine are transmitted and received via the communication bus, vibrations may still occur in the electric vehicle even if a corrected command value according to the vehicle characteristics is used.

[0005] The present invention aims to provide a control method for an electric vehicle that can better suppress vibrations of the electric vehicle through vibration damping control when controlling a rotating electric machine using a first electronic control unit and a second electronic control unit that communicates with the first electronic control unit via a communication bus, and a control device for the electric vehicle.

[0006] FIG. 1 is a block diagram showing a schematic configuration of an electric vehicle. FIG. 2 is a block diagram showing the configurations of a first ECU and a second ECU. FIG. 3 is a block diagram showing the configuration of a quantization calculation unit. FIG. 4 is a block diagram showing the configuration of a noise shaping quantization calculation unit. FIG. 5 is a graph showing the time change of an upper limit torque command value before quantization. FIG. 6 is a graph showing the time change of longitudinal acceleration in a first comparative example. FIG. 7 is a graph showing the time change of longitudinal acceleration in a second comparative example. FIG. 8 is a graph showing the error of a final torque command value and its amplitude spectrum. FIG. 9 is a graph showing the change of longitudinal acceleration. FIG. 10 is a graph showing the error of longitudinal acceleration and its amplitude spectrum. FIG. 11 is a block diagram showing the configuration of a noise shaping quantization calculation unit according to a second embodiment. FIG. 12 is a graph showing the error of a final torque command value and its amplitude spectrum. FIG. 13 is a graph showing the time change of longitudinal acceleration. FIG. 14 is a graph showing the error of longitudinal acceleration and its amplitude spectrum. FIG. 15 is a block diagram showing the configuration of a noise shaping quantization calculation unit according to a third embodiment. FIG. 16 is a graph showing the error of a final torque command value and its amplitude spectrum. Fig. 17 is a graph showing changes in longitudinal acceleration over time. Fig. 18 is a graph showing errors in longitudinal acceleration and their amplitude spectra. Fig. 19 is a block diagram showing the configuration of a noise shaping quantization calculation unit according to a fourth embodiment.

[0007] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0008] 1 is a block diagram showing a schematic configuration of an electric vehicle 100. As shown in FIG. 1, the electric vehicle 100 includes a battery 10, a rotating electric machine 11, an inverter 12, and a first ECU 13.

[0009] The battery 10 is a DC power supply that supplies power to each part of the electric vehicle 100, such as the rotating electric machine 11. The battery 10 is formed of, for example, a lithium-ion battery and is rechargeable. When the temperature of the battery 10 is low or when the charging rate (SOC: State of Charge) of the battery 10 is low, the power that the battery 10 can output (hereinafter referred to as outputtable power) decreases.

[0010] The rotating electric machine 11 is an electric motor that functions as a drive source for the electric vehicle 100, or a generator that constitutes an on-board power generation system. In this embodiment, the rotating electric machine 11 is a three-phase AC synchronous motor that functions as a drive source for the electric vehicle 100. Therefore, the rotating electric machine 11 is connected to drive wheels via a differential gear, a drive shaft, etc. (hereinafter referred to as a drive system). The rotating electric machine 11 is connected to a drive wheel ... For example, the accelerator operation amount A ｐｏ The torque generated by the rotary electric machine 11 is transmitted to the drive wheels via the drive system. ｐｏ A driving force or braking force corresponding to the

[0011] The inverter 12 converts DC power supplied from the battery 10 into AC power and supplies it to the rotating electric machine 11. In this way, the inverter 12 controls the operation of the rotating electric machine 11. Furthermore, when the rotating electric machine 11 is rotated by a driving wheel or the like, the inverter 12 can convert the AC power generated by the rotating electric machine 11 into DC power to charge the battery 10.

[0012] The first ECU 13 is a higher-level electronic control unit (ECU) that comprehensively controls each part of the electric vehicle 100. When the electric vehicle 100 is a hybrid vehicle, the first ECU 13 is, for example, a so-called HEVC (Hybrid Electric Vehicle Controller). The first ECU 13 calculates basic command values ​​(hereinafter referred to as basic command values) that determine the operating state of the rotating electric machine 11 in relation to the control of the rotating electric machine 11.

[0013] In this embodiment, the first ECU 13 controls the accelerator operation amount A ｐｏ and the electrical angular velocity ω of the rotating electrical machine 11 ｅ Based on this, the required torque command value T ｒｅｑ ＊ The required torque command value T ｒｅｑ ＊is a command value for the torque that the rotary electric machine 11 should generate in order to generate the driving force or braking force required by the operation of the accelerator. ｒｅｑ ＊ is one of the basic command values ​​related to the control of the rotating electrical machine 11.

[0014] Accelerator operation amount A ｐｏ is appropriately acquired using an accelerator opening sensor (not shown). ｅ is appropriately acquired using a rotation sensor 14 provided in the rotating electric machine 11. Furthermore, the first ECU 13 acquires an electrical angular velocity ω as a parameter representing the rotation state of the rotating electric machine 11. ｅ Instead of [rad / s], mechanical angular velocity ω ｍ [rad / s] or rotation speed N ｍ In this case, the first ECU 13 uses the rotation sensor 14 to detect the mechanical angular velocity ω of the rotating electrical machine 11. ｍ and rotation speed N ｍ Get.

[0015] Required torque command value T ｒｅｑ ＊ In addition, the first ECU 13 determines the upper limit torque command value T ｌｉｍ ＊ The upper limit torque command value T ｌｉｍ ＊ is a command value that determines the upper limit of the torque of the rotating electrical machine 11. ｌｉｍ ＊ is one of the basic command values ​​related to the control of the rotating electrical machine 11. Specifically, the first ECU 13 determines, for example, the limit power P ｌｉｍ and the electrical angular velocity ω of the rotating electrical machine 11 ｅ Using the upper limit torque command value T ｌｉｍ ＊ Calculate the following.

[0016] Limited power P ｌｉｍ is the power that can be substantially supplied from the battery 10 to the inverter 12. ｌｉｍ is set according to the available power output of the battery 10. Therefore, when the temperature or SOC of the battery 10 decreases, the limit power P ｌｉｍAs a result, when the temperature or SOC of the battery 10 decreases, the upper limit torque command value T ｌｉｍ ＊ In this embodiment, the first ECU 13 receives the limited power P ｌｉｍ However, the first ECU 13 may include a battery controller. In this case, the first ECU 13 determines the limit power P by itself based on the temperature and SOC of the battery 10. ｌｉｍ can be set.

[0017] As described above, the required torque command value T ｒｅｑ ＊ and the upper limit torque command value T ｌｉｍ ＊ is input to the inverter 12. The inverter 12 then outputs the required torque command value T ｒｅｑ ＊ and the upper limit torque command value T ｌｉｍ ＊ Based on this, power is supplied from the battery 10 to the rotating electrical machine 11.

[0018] In this embodiment, the inverter 12 includes a second ECU 15 and a bridge circuit 16 .

[0019] The second ECU 15 is an electronic control unit (ECU) subordinate to the first ECU 13. The second ECU 15 controls the operation of the rotating electric machine 11 by driving the bridge circuit 16 in accordance with the basic command value calculated by the first ECU 13. In particular, the second ECU 15 uses the basic command value to calculate a correction command value that suppresses vibrations of the electric vehicle 100. The second ECU 15 then drives the inverter 12 based on the correction command value, thereby controlling the operation of the rotating electric machine 11 so as to suppress vibrations of the electric vehicle 100. In other words, the second ECU 15 is an electronic control unit responsible for vibration suppression control.

[0020] In this embodiment, the second ECU 15 calculates the required torque command value T ｒｅｑ ＊ and the upper limit torque command value T ｌｉｍ ＊The bridge circuit 16 is driven in accordance with the above, thereby controlling the operation of the rotating electric machine 11 .

[0021] Specifically, the second ECU 15 calculates the required torque command value T ｒｅｑ ＊ The vibration damping torque command value T ｄａｍ ＊ The vibration damping torque command value T ｄａｍ ＊ is set to the required torque command value T ｒｅｑ ＊ is a torque command value (corrected command value) obtained by correcting the torque command value T in accordance with the vehicle characteristics of the electric vehicle 100. ｄａｍ ＊ By controlling the rotating electric machine 11 based on the required torque command value T ｒｅｑ ＊ While outputting torque according to the torque, vibration of the electric vehicle 100 is suppressed.

[0022] The second ECU 15 calculates the vibration damping torque command value T ｄａｍ ＊ When calculating the required torque command value T ｒｅｑ ＊ is the upper limit torque command value T ｌｉｍ ＊ When it is larger than (T ｒｅｑ ＊ >T ｌｉｍ ＊ ), required torque command value T ｒｅｑ ＊ is the upper limit torque command value T ｌｉｍ ＊ The second ECU 15 also limits the calculated vibration damping torque command value T ｄａｍ ＊ is the upper limit torque command value T ｌｉｍ ＊ When it is larger than (T ｄａｍ ＊ >T ｌｉｍ ＊ ), vibration damping torque command value T ｄａｍ ＊ is the upper limit torque command value T ｌｉｍ ＊ Restrict based on.

[0023] The second ECU 15 calculates the vibration damping torque command value T ｄａｍ ＊ , DC voltage V input from battery 10 ｄｃ , and the currents flowing through the UVW phases of the rotating electrical machine 11 (hereinafter referred to as three-phase currents i ｕｖｗ ) to control the on / off of the switching elements that make up the bridge circuit 16. As a result, the rotating electrical machine 11 controls the damping torque command value T ｄａｍ ＊ The torque generated corresponds to the DC voltage V of the battery 10. ｄｃ is suitably acquired by the voltage sensor 17. In addition, the three-phase current i ｕｖｗ is suitably acquired by the current sensor 18.

[0024] Bridge circuit 16 includes a pair of switching elements for each of the UVW phases. These switching elements are switched on / off in accordance with a so-called PWM (Pulse Width Modulation) signal. Bridge circuit 16 controls the oscillation-damping torque command value T ｄａｍ ＊ In order to realize a torque output according to the three-phase current i ｕｖｗ Control.

[0025] In this embodiment, since the rotating electric machine 11 is an electric motor that functions as a drive source for the electric vehicle 100, the torque of the rotating electric machine 11 is controlled so that the torque matches or follows a predetermined torque. For this reason, the first ECU 13 uses the required torque command value T ｒｅｑ ＊ and the upper limit torque command value T ｌｉｍ ＊ The second ECU 15 calculates the damping torque command value T ｄａｍ ＊ However, when the rotating electric machine 11 is a generator of an on-board power generation system, the rotating electric machine 11 has a rotation speed (number of revolutions N ｍThe torque is controlled so that the rotational speed command value, etc., coincides with or follows a predetermined rotational speed. Therefore, when the rotating electrical machine 11 is a generator, the first ECU 13 can calculate a rotational speed command value, etc. as the basic command value. In this case, the second ECU 15 calculates a vibration-damping rotational speed command value as the correction command value in response to the basic command value.

[0026] As described above, the control device (controller) of the electric vehicle 100 is composed of a plurality of electronic control units (computers) including the first ECU 13 and the second ECU 15. The electronic control units thus provided in a distributed manner are connected to each other via a so-called CAN (Controller Area Network) so as to be able to communicate with each other.

[0027] For example, the first ECU 13 and the second ECU 15 transmit and receive basic command values ​​via the CAN communication bus 19. Therefore, the first ECU 13 discretizes the calculated basic command values ​​for transmission via the CAN. Specifically, the first ECU 13 samples the calculated basic command values ​​in accordance with the transmission interval of the CAN, quantizes them with a quantization width according to the constraints of the CAN communication volume, and transmits the discretized values ​​to the second ECU 15.

[0028] In summary, the first ECU 13 is configured to calculate a basic command value that determines the operating state of the rotating electric machine 11, quantize the basic command value with a quantization width that corresponds to the constraints on the communication volume on the communication bus 19, and transmit the quantized basic command value, which is the quantized basic command value, to the second ECU 15 via the communication bus 19.

[0029] In the following, when a distinction is necessary, the required torque command value T before quantization, which has a substantially continuous value, is used. ｒｅｑ ＊ The continuous required torque command value T ｒｅｑ－ｃｏｎｔ ＊ and the quantized required torque command value T ｒｅｑ ＊ quantized required torque command value T ｒｅｑ－ｑｕａｎｔ ＊ Similarly, the upper limit torque command value T ｌｉｍ ＊ The continuous upper limit torque command value T ｌｉｍ－ｃｏｎｔ＊ and the quantized upper limit torque command value T ｌｉｍ ＊ The quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ The continuous required torque command value T ｒｅｑ－ｃｏｎｔ ＊ and the continuous upper limit torque command value T ｌｉｍ－ｃｏｎｔ ＊ is the basic command value before quantization, and the quantized required torque command value T ｒｅｑ－ｑｕａｎｔ ＊ and the quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ is a quantization basic command value. A substantially continuous value refers to an actual continuous value or a discrete value with a quantization width smaller than the quantization width of CAN communication.

[0030] The second ECU 15 also receives the quantized basic command value and performs vibration damping control using the quantized basic command value.

[0031] In summary, the second ECU 15 is configured to receive a quantized basic command value, which is a quantized basic command value, from the first ECU 13, use the quantized basic command value to calculate a corrected command value that suppresses vibrations of the electric vehicle 100, and control the operation of the rotating electric machine 11 based on the corrected command value.

[0032] When the basic command value quantized for CAN communication is used to calculate the correction command value for vibration suppression control, vibrations may occur in the electric vehicle 100, even though the correction command value for vibration suppression is used. That is, when the basic command value for vibration suppression control is quantized for communication between the first ECU 13 and the second ECU 15, vibrations in the electric vehicle 100 may not be sufficiently suppressed and may remain even if vibration suppression control is performed. When the rotating electric machine 11 is an electric motor, typically, the required torque command value T ｒｅｑ ＊ or vibration damping torque command value T ｄａｍ ＊ is the upper limit torque command value T ｌｉｍ ＊ When the vibration damping torque command value T ｄａｍ ＊Even when vibration damping control using the above is performed, vibrations may still occur in the electric vehicle 100. Therefore, in this embodiment, when the first ECU 13 quantizes the basic command value, the first ECU 13 shapes the quantization noise so that a predetermined frequency component including a resonance frequency of vibrations occurring in the electric vehicle 100 moves toward a higher frequency. That is, when the first ECU 13 quantizes the basic command value, the first ECU 13 performs noise shaping (frequency shaping) to shape the frequency of the quantization noise.

[0033] Fig. 2 is a block diagram showing the configuration of the first ECU 13 and the second ECU 15. As shown in Fig. 2, the first ECU 13 includes, for example, a required torque calculation unit 21, an upper limit torque calculation unit 22, a quantization calculation unit 23, and a noise shaping quantization calculation unit 24.

[0034] The required torque calculation unit 21 calculates the accelerator operation amount A ｐｏ and electrical angular velocity ω ｅ Based on this, the continuous required torque command value T ｒｅｑ－ｃｏｎｔ ＊ Specifically, the first ECU 13 calculates the accelerator operation amount A ｐｏ and electrical angular velocity ω ｅ and the continuous required torque command value T ｒｅｑ－ｃｏｎｔ ＊ Therefore, the first ECU 13 can calculate the accelerator operation amount A by referring to this torque map. ｐｏ and electrical angular velocity ω ｅ Continuously required torque command value T ｒｅｑ－ｃｏｎｔ ＊ Calculate the following.

[0035] The upper limit torque calculation unit 22 calculates the limit power P ｌｉｍ and electrical angular velocity ω ｅ Based on this, the continuous upper limit torque command value T ｌｉｍ－ｃｏｎｔ ＊ Specifically, the first ECU 13 calculates the limit power P ｌｉｍ is the electrical angular velocity ω ｅ By dividing by this, the continuous upper limit torque command value T ｌｉｍ－ｃｏｎｔ ＊ Calculate the following.

[0036]

[0037] The quantization calculation unit 23 calculates the continuous required torque command value T ｒｅｑ－ｃｏｎｔ ＊ is quantized and transmitted via the communication bus 19 as a quantized required torque command value T ｒｅｑ－ｑｕａｎｔ ＊ to the second ECU 15.

[0038] The noise shaping quantization calculation unit 24 calculates the continuous upper limit torque command value T ｌｉｍ－ｃｏｎｔ ＊ is quantized and transmitted via the communication bus 19 as the quantized upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ to the second ECU 15. The noise shaping quantization calculation unit 24 also transmits the continuous upper limit torque command value T ｌｉｍ－ｃｏｎｔ ＊ When quantizing the quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ Shapes the quantization noise superimposed on

[0039] The second ECU 15 includes, for example, a first limiter 25 , a vibration damping calculation unit 26 , a second limiter 27 , and a torque control unit 28 .

[0040] The first limiter 25 receives the quantized required torque command value T ｒｅｑ－ｑｕａｎｔ ＊ and the quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ and, if necessary, the quantized required torque command value T ｒｅｑ－ｑｕａｎｔ ＊ The quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ Specifically, the first limiter 25 limits the quantized required torque command value T ｒｅｑ－ｑｕａｎｔ ＊ is the quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ If it exceeds (T ｒｅｑ－ｑｕａｎｔ ＊ >T ｌｉｍ－ｑｕａｎｔ＊ ), quantized required torque command value T ｒｅｑ－ｑｕａｎｔ ＊ The quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ On the other hand, the quantized required torque command value T ｒｅｑ－ｑｕａｎｔ ＊ is the quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ If T ｒｅｑ－ｑｕａｎｔ ＊ ≦T ｌｉｍ－ｑｕａｎｔ ＊ ), the first limiter 25 essentially limits the quantized required torque command value T ｒｅｑ－ｑｕａｎｔ ＊ As a result, the first limiter 25 does not limit the quantized required torque command value T ｒｅｑ－ｑｕａｎｔ ＊ or the quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ is the limited required torque command value T ｒｅｑ－ｌｉｍ ＊ Output as

[0041] The vibration suppression calculation unit 26 calculates the limit required torque command value T ｒｅｑ－ｌｉｍ ＊ is the electrical angular velocity ω ｅ and corrected based on the vehicle characteristics of the electric vehicle 100, the vibration damping torque command value T ｄａｍ ＊ Specifically, the vibration damping calculation unit 26 calculates the electrical angular velocity ω based on the torque transmission characteristics of the electric vehicle 100. ｅ Then, the vibration suppression calculation unit 26 calculates the torque of the rotating electrical machine 11 from the limited required torque command value T ｒｅｑ－ｌｉｍ ＊ and the torque to be generated by the input of ｅ The difference between these is set to the limited required torque command value T ｒｅｑ－ｌｉｍ ＊ The limited required torque command value T ｒｅｑ－ｌｉｍ ＊ By compensating (correcting) the vibration damping torque command value T ｄａｍ ＊ Calculate the following.

[0042] The second limiter 27 adjusts the vibration damping torque command value T ｄａｍ＊ The quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ Specifically, the second limiter 27 limits the vibration damping torque command value T ｄａｍ ＊ is the quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ If it exceeds (T ｄａｍ ＊ >T ｌｉｍ－ｑｕａｎｔ ＊ ), vibration damping torque command value T ｄａｍ ＊ The quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ On the other hand, the vibration suppression torque command value T ｄａｍ ＊ is the quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ If T ｄａｍ ＊ ≦T ｌｉｍ－ｑｕａｎｔ ＊ ), the second limiter 27 substantially limits the quantized required torque command value T ｒｅｑ－ｑｕａｎｔ ＊ As a result, the second limiter 27 does not limit the vibration damping torque command value T ｄａｍ ＊ or the quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ is the final torque command value T ｆｉｎ ＊ Output as

[0043] The torque control unit 28 calculates the final torque command value T ｆｉｎ ＊ , the DC voltage V of the battery 10 ｄｃ , and three-phase current i ｕｖｗ The torque control unit 28 generates a PWM signal for controlling the bridge circuit 16 based on the PWM signal. The torque control unit 28 controls the rotating electric machine 11 by switching on / off each switching element of the bridge circuit 16 in accordance with the PWM signal. As a result, the rotating electric machine 11 controls the final torque command value T ｆｉｎ ＊ The torque output is in accordance with the

[0044] 3 is a block diagram showing the configuration of the quantization calculation unit 23. As shown in FIG. 3, the required torque command value Tｒｅｑ ＊ (Continuously required torque command value T ｒｅｑ－ｃｏｎｔ ＊ The quantization operation unit 23 for quantizing the quantization signal ) is configured by a uniform quantizer 31.

[0045] The uniform quantizer 31 is a static quantizer that quantizes the input u by rounding, truncation, or other fixed method. The uniform quantizer 31 is expressed by a mapping q shown in the following equation (2) using the number of bits N and the quantization width d. "R" is a set of real numbers. As shown in the following equation (3), the input u and output v of the uniform quantizer 31 are real numbers. Therefore, the input / output relationship of the uniform quantizer 31 is expressed by the following equation (4). Furthermore, as shown in the following equation (5), the quantization noise ω (quantization error) is expressed by the deviation between the output v and the input u. Therefore, the quantization noise ω by the uniform quantizer 31 depends on the input u. For example, the larger the change in the input u, the higher the frequency of the quantization noise ω will be distributed, and the smaller the change in the input u, the lower the frequency of the quantization noise ω will be concentrated. Note that FIG. 3 shows the Z-transformed input u(z) and output v(z). In the following, the quantization noise ω will also be expressed as a Z-transformed value (ω(z)). In this embodiment, the input u(z) of the uniform quantizer 31 is the continuous required torque command value T ｒｅｑ－ｃｏｎｔ ＊ and the output v(z) is the quantized required torque command value T ｒｅｑ－ｑｕａｎｔ ＊ is.

[0046]

[0047] 4 is a block diagram showing the configuration of the noise shaping quantization calculation unit 24. As shown in FIG. ｌｉｍ ＊ (Continuous upper limit torque command value T ｌｉｍ－ｃｏｎｔ ＊ The noise shaping quantization operation unit 24 that quantizes the signal σ is made up of a uniform quantizer 32, a quantization noise estimation unit 33, a compensation value operation unit 34, and a compensation unit 35.

[0048] The uniform quantizer 32 is configured similarly to the uniform quantizer 31 .

[0049] However, the amount of communication in the CAN, i.e., the amount of data in a CAN packet (frame), is limited. ｒｅｑ ＊ or vibration damping torque command value T ｄａｍ ＊ is the upper limit torque command value T ｌｉｍ ＊ Therefore, the number of bits N and the quantization width d of the uniform quantizer 32 are set to values ​​different from the number of bits N and the quantization width d of the uniform quantizer 31 so as to satisfy the constraints on the communication volume of the CAN.

[0050] Specifically, the required torque command value T ｒｅｑ ＊ The number of bits N and the quantization width d of the uniform quantizer 31 are N ｒｅｑ , d ｒｅｑ and the upper limit torque command value T ｌｉｍ ＊ The number of bits N and the quantization width d of the uniform quantizer 32 are N ｌｉｍ , d ｌｉｍ Then, N ｒｅｑ >N ｌｉｍ , d ｒｅｑ <d ｌｉｍ That is, the upper limit torque command value T ｌｉｍ ＊ The number of bits N (N ｌｉｍ ) is the required torque command value T ｒｅｑ ＊ The number of bits N (N ｒｅｑ ) and the upper limit torque command value T ｌｉｍ ＊ quantization width d (d ｌｉｍ ) is the required torque command value T ｒｅｑ ＊ quantization width d (d ｒｅｑ ) is larger than the upper limit torque command value T ｌｉｍ ＊ is the required torque command value T ｒｅｑ ＊ This means that the quantization is more coarse than

[0051] Therefore, the upper limit torque command value T ｌｉｍ ＊ The quantization noise ω(z) superimposed on the required torque command value Tｒｅｑ ＊ Therefore, the required torque command value T ｒｅｑ ＊ or vibration damping torque command value T ｄａｍ ＊ is the upper limit torque command value T ｌｉｍ ＊ When vibration suppression control is performed in a state limited to ｄａｍ ＊ Even if the upper limit torque command value T ｌｉｍ ＊ The quantized noise ω(z) superimposed on the signal ω(z) may cause vibrations in the electric vehicle 100.

[0052] Therefore, the noise shaping quantization calculation unit 24 calculates the upper limit torque command value T ｌｉｍ ＊ The quantization noise ω(z) superimposed on the signal is shaped.

[0053] In this embodiment, the noise shaping quantization calculation unit 24 is configured by a dynamic quantizer in which a quantization noise estimation unit 33 , a compensation value calculation unit 34 , and a compensation unit 35 are added to a uniform quantizer 32 .

[0054] The quantization noise estimation unit 33 calculates the upper limit torque command value T ｌｉｍ ＊ Specifically, the quantization noise estimation unit 33 calculates the quantization noise ω(z) by calculating the difference between the input u(z) and the output v(z) of the uniform quantizer 32. In this embodiment, the quantization noise estimation unit 33 calculates the quantization noise ω(z) by subtracting the input u(z) from the output v(z) of the uniform quantizer 32. That is, ω(z) = v(z) - u(z). The input u(z) of the uniform quantizer 32 is calculated based on the continuous upper limit torque command value T ｌｉｍ－ｃｏｎｔ ＊ and the output v(z) is the quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ is.

[0055] The compensation value calculation unit 34 calculates a compensation value for the input u(z) of the uniform quantizer 32 based on the quantization noise ω(z). In this embodiment, the compensation value calculation unit 34 includes a filter F that extracts predetermined frequency components including the resonance frequency of vibrations occurring in the electric vehicle 100. ω (z). Therefore, the compensation value is F ω (z)ω(z).

[0056] The compensation unit 35 compensates the output v(z) of the noise shaping quantization operation unit 24 by subtracting a compensation value from the input u(z) to the uniform quantizer 32. Therefore, the output v(z) of the noise shaping quantization operation unit 24 is expressed by the following equation (6).

[0057]

[0058] In this embodiment, the filter F ω (z) is a low-pass filter (LPF) expressed by the following equation (7) using parameters A and B. Note that a low-pass filter is usually a delay element. However, as shown in the above equation (6), the filter F ω (z) does not substantially act on the input u(z) but acts only on the quantization noise ω(z). Therefore, in the noise shaping quantization operation unit 24, the output v(z) is not delayed with respect to the input u(z).

[0059]

[0060] The parameters A and B are used to calculate the filter F ω The cutoff frequency of (z) is set in advance to be equal to or higher than the resonant frequency of vibrations occurring in electric vehicle 100. However, the sum of parameters A and B is 1 (A+B=1). Furthermore, parameter B in the denominator needs to satisfy the condition |B|<1 for stability.

[0061] When the parameter B is changed from +1 to −1 so that |B|<1 is satisfied under the condition of A+B=1, the filter F ω In other words, if the parameter B is increased, the cutoff frequency of the filter F ωThe cutoff frequency of (z) becomes larger, and if the parameter B is made smaller, the filter F ω The cutoff frequency of (z) becomes smaller. Therefore, the parameters A and B are ω The cutoff frequency of (z) can be set by adaptation so as to be equal to or higher than the resonant frequency of vibrations occurring in the electric vehicle 100.

[0062] In addition, the filter F ω The cutoff frequency of (z) has a maximum value. The maximum value of the cutoff frequency is the sampling period τ ｓ However, in many cases, the resonant frequency of the vibration occurring in the electric vehicle 100 is in a low frequency band of the order of 10 Hz or less. For this reason, the parameters A and B are usually ω The cutoff frequency of (z) can be set to be equal to or higher than the resonant frequency of vibrations occurring in the electric vehicle 100 .

[0063] In this embodiment, the parameter A is calculated based on the sampling period τ ｓ and time constant τ ｃ The ratio τ ｓ / τ ｃ The parameter B is expressed by the time constant τ ｃ and sampling period τ ｓ and the time constant τ ｃ and the time constant τ ｃ is the filter F ω The cutoff frequency of (z) is determined in accordance with the vehicle characteristics of the electric vehicle 100 so as to be equal to or higher than the resonant frequency of vibrations occurring in the electric vehicle 100 .

[0064]

[0065] As mentioned above, the filter F ωSince (z) is a low-pass filter and the compensation unit 35 subtracts it from the input u(z), the noise shaping quantization calculation unit 24 as a whole acts like a high-pass filter (HPF) on the quantization noise ω(z). However, when the quantization noise ω(z) below the cutoff frequency is reduced in this manner, the gain increases in the high-frequency band above the cutoff frequency. In other words, the quantization noise ω(z) is essentially shaped so that a predetermined frequency component including the resonant frequency of vibrations occurring in the electric vehicle 100 moves toward the high-frequency side (the frequency band above the cutoff frequency). However, the quantization noise ω(z) belonging to the high-frequency band above the cutoff frequency is far from the resonant frequency and therefore hardly causes vibrations occurring in the electric vehicle 100.

[0066] Therefore, the required torque command value T ｒｅｑ ＊ or vibration damping torque command value T ｄａｍ ＊ However, the upper limit torque command value T ｌｉｍ ＊ Even when the upper limit torque command value T ｌｉｍ ＊ By shaping the quantized noise ω(z) superimposed on the vibration of the electric vehicle 100, the vibration of the electric vehicle 100 can be more effectively suppressed by the vibration suppression control.

[0067] The vibration damping effect in the electric vehicle 100 configured as described above will now be described.

[0068] FIG. 5 shows the upper limit torque command value T ｌｉｍ ＊ (Continuous upper limit torque command value T ｌｉｍ－ｃｏｎｔ ＊ 5 is a graph showing the time change of the continuous upper limit torque command value T ｌｉｍ－ｃｏｎｔ ＊ is the time t １ rises at time t ２ In the following, the required torque command value T ｒｅｑ ＊ and the vibration damping torque command value T ｄａｍ ＊ is the upper limit torque command value Tｌｉｍ ＊ is limited to the upper limit torque command value T ｌｉｍ ＊ A scene in which the rotating electrical machine 11 is controlled in accordance with the above will be described.

[0069] FIG. 6 shows the longitudinal acceleration A Ｌ 1 is a graph showing the time change of the upper limit torque command value T ｌｉｍ＊ is simply quantized by the uniform quantizer 32, and the quantization noise ω(z) is not shaped. That is, the first comparative example is an example in which the noise-shaping quantization calculation unit 24 of this embodiment is configured with only the uniform quantizer 32.

[0070] FIG. 6A shows the longitudinal acceleration A occurring in the electric vehicle 100 over the time range corresponding to FIG. 5 . Ｌ 6(A) and 6(B) show the longitudinal acceleration A of the first comparative example. Ｌ is shown by a solid line, and the ideal longitudinal acceleration A Ｌ The change in the ideal longitudinal acceleration A Ｌ The change in the continuous upper limit torque command value T ｌｉｍ－ｃｏｎｔ ＊ The longitudinal acceleration A when the rotating electric machine 11 is controlled according to Ｌ This is a change.

[0071] As shown in FIGS. 6A and 6B, the upper limit torque command value T ｌｉｍ ＊ When quantizing ２ Hereafter, the upper limit torque command value T ｌｉｍ ＊ When the longitudinal acceleration A of the electric vehicle 100 gradually decreases, Ｌ That is, the upper limit torque command value T ｌｉｍ ＊ If the quantization noise ω(z) is not shaped, the torsional vibration of the drive shaft will cause the electric vehicle 100 to vibrate back and forth.

[0072] FIG. 7 shows the longitudinal acceleration AＬ 10 is a graph showing the time change of the upper limit torque command value T ｌｉｍ＊ is simply quantized by the uniform quantizer 32, and the second ECU 15 converts the received quantized upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ is smoothed by a low-pass filter and used for vibration suppression control. ｌｉｍ－ｑｕａｎｔ ＊ The time constant of the low-pass filter that smooths the signal is adjusted in advance so as to reduce a predetermined frequency band that includes the resonance frequency of vibrations that occur in the electric vehicle 100.

[0073] FIG. 7A shows the longitudinal acceleration A occurring in the electric vehicle 100 over the time range corresponding to FIG. 5 . Ｌ 7(A) and 7(B) show the longitudinal acceleration A of the second comparative example. Ｌ is shown by a solid line, and the ideal longitudinal acceleration A Ｌ The change in is shown by the dashed line.

[0074] As shown in FIGS. 7A and 7B, the upper limit torque command value T ｌｉｍ ＊ is smoothed by an LPF in the receiving side second ECU 15 and used for vibration suppression control, the longitudinal acceleration A Ｌ vibration is reduced, but the longitudinal acceleration A Ｌ The vibration of the ideal longitudinal acceleration A Ｌ There is a delay in response.

[0075] In this second comparative example, the quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ If the time constant of the low-pass filter that smooths the longitudinal acceleration A is increased, Ｌ Therefore, the torque limit is delayed and the output power of the battery 10 is reduced to the limited power P ｌｉｍ On the other hand, the quantization upper limit torque command value Tｌｉｍ－ｑｕａｎｔ ＊ The delay can be improved by reducing the time constant of the low-pass filter that smooths the longitudinal acceleration A Ｌ More vibration remains.

[0076] That is, the upper limit torque command value T ｌｉｍ ＊ When the quantization noise ω(z) is reduced by a low-pass filter on the receiving side, the longitudinal vibration of the electric vehicle 100 caused by the quantization noise ω(z) cannot be suppressed completely. Furthermore, if the longitudinal vibration is to be sufficiently suppressed, the output power of the battery 10 must be reduced to the limit power P ｌｉｍ There is a risk of exceeding the limit.

[0077] FIG. 8 shows the final torque command value T ｆｉｎ ＊ 10 is a graph showing the error ΔT of the final torque command value T ｆｉｎ ＊ The error ΔT is transmitted via the CAN to the quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ The final torque command value T ｆｉｎ ＊ and the continuous upper limit torque command value T ｌｉｍ－ｃｏｎｔ ＊ The ideal final torque command value T ｆｉｎ ＊ This is the difference between and.

[0078] FIG. 8A shows the final torque command value T ｆｉｎ ＊ 8B is a graph showing the change over time of the error ΔT in the first comparative example. ｆｉｎ 8B is a graph showing the amplitude spectrum of the error ΔT. The horizontal axis of FIG. 8B is the frequency f [Hz], and the vertical axis is the amplitude |P １ (f)|.

[0079] FIG. 8C shows the final torque command value T ｆｉｎ ＊ 8D is a graph showing the change over time of the error ΔT in this embodiment. ｆｉｎ 8(D) is a graph showing the amplitude spectrum of the frequency f [Hz], and the vertical axis is the error ΔTｆｉｎ Amplitude of |P １ (f)|.

[0080] As shown in FIG. 8(A) and FIG. 8(C), the upper limit torque command value T ｌｉｍ ＊ In this embodiment, the upper limit torque command value T ｌｉｍ ＊ is simply quantized by the uniform quantizer 32, the final torque command value T ｆｉｎ ＊ The error ΔT increases.

[0081] 8B and 8D, in the first comparative example, the components of the error ΔT are concentrated in the low frequency band of 10 Hz or less, whereas in this embodiment, the low frequency components of the error ΔT are almost eliminated and the high frequency components of 10 Hz or more increase. This is because the noise shaping quantization calculation unit 24 shifts the low frequency components of the quantization noise ω(z) to the high frequency side by noise shaping.

[0082] As described above, the resonant frequency of vibrations occurring in the electric vehicle 100 is in a low frequency band on the order of 10 Hz or less, and therefore, as described above, the noise shaping quantization calculation unit 24 shifts the low frequency components of the quantization noise ω(z) to the high frequency side, thereby better suppressing vibrations in the electric vehicle 100. Specifically, this is as follows.

[0083] FIG. 9 shows the longitudinal acceleration A Ｌ 10 is a graph showing the change in

[0084] FIG. 9A shows the longitudinal acceleration A of the electric vehicle 100 in the first comparative example for the time range corresponding to FIG. 5 . Ｌ 9(A) and 9(B) show the longitudinal acceleration A of the first comparative example. Ｌ is shown by a solid line, and the ideal longitudinal acceleration A Ｌ The change in is shown by the dashed line.

[0085] FIG. 9C shows the longitudinal acceleration A occurring in the electric vehicle 100 in this embodiment for the time range corresponding to FIG. 5 . Ｌ 9(C) and 9(D) show the change over time of the longitudinal acceleration A of this embodiment. Also, FIG. 9(D) is an enlarged view of the range indicated by the dashed line in FIG. 9(C). The range indicated by the dashed line in FIG. 9(C) is the same as the range indicated by the dashed line in FIG. 9(A). In FIGS. 9(C) and 9(D), the longitudinal acceleration A of this embodiment is Ｌ is shown by a solid line, and the ideal longitudinal acceleration A Ｌ The change in is shown by the dashed line.

[0086] As shown in FIGS. 9A to 9D, when the first comparative example is compared with this embodiment, in this embodiment, the time t ２ Hereafter, the upper limit torque command value T ｌｉｍ ＊ When the longitudinal acceleration A of the electric vehicle 100 gradually decreases, Ｌ In this embodiment, the vibration occurring in the longitudinal acceleration A Ｌ There is no delay.

[0087] FIG. 10 shows the longitudinal acceleration A Ｌ Error ΔA Ｌ 10 is a graph showing the longitudinal acceleration A and its amplitude spectrum. Ｌ Error ΔA Ｌ is transmitted via the CAN to the quantization upper limit torque command value T ｌｉｍ－ｑｕａｎｔ ＊ When using the longitudinal acceleration A Ｌ and the continuous upper limit torque command value T ｌｉｍ－ｃｏｎｔ ＊ Ideal longitudinal acceleration A Ｌ This is the difference between and.

[0088] FIG. 10A shows the longitudinal acceleration A Ｌ Error ΔA Ｌ 10B is a graph showing the change over time of the error ΔA Ｌ 10B is a graph showing the amplitude spectrum of the frequency f [Hz], and the vertical axis is the error ΔA Ｌ Amplitude of |P ２ (f)|.

[0089] FIG. 10C shows the longitudinal acceleration A Ｌ Error ΔA Ｌ 10D is a graph showing the time change of the error ΔA Ｌ 10(D) is a graph showing the amplitude spectrum of the frequency f [Hz], and the vertical axis is the error ΔA Ｌ Amplitude of |P ２ (f)|.

[0090] As shown in FIG. 10(A) and FIG. 10(C), the upper limit torque command value T ｌｉｍ ＊ In this embodiment, the upper limit torque command value T ｌｉｍ ＊ The longitudinal acceleration A Ｌ Error ΔA Ｌ This is also apparent from Figures 9(B) and 9(D).

[0091] Furthermore, as shown in FIGS. 10B and 10D, in the first comparative example, the error ΔA Ｌ The component of is widely distributed in the low frequency band below 10 Hz, whereas in this embodiment, the error ΔA Ｌ is slightly distributed in a frequency band of about 10 Hz, that is, in the vicinity of the resonance frequency of the vibrations occurring in the electric vehicle 100. Therefore, as in this embodiment, the noise shaping quantization calculation unit 24 moves the low frequency components of the quantization noise ω(z) to the high frequency side, thereby better suppressing the vibrations of the electric vehicle 100.

[0092] In the first embodiment, the filter F constituting the compensation value calculation unit 34 ω Although an example in which (z) is a low-pass filter has been described, the present invention is not limited to this. ω It is sufficient that the filter F (z) has the characteristic of extracting a predetermined frequency component including the resonance frequency of the vibration generated in the electric vehicle 100. ω (z) can also be constituted by, for example, a band-pass filter.

[0093] Second Embodiment In the first embodiment, the compensation value calculation unit 34 calculates F ω (z) = Az －１ / (1-Bz －１ ) form, but is not limited to this. The noise shaping quantization calculation section 24 may be configured by a dynamic quantizer of another form.

[0094] 11 is a block diagram showing the configuration of the noise shaping quantization calculation unit 24 according to the second embodiment. As shown in FIG. 11, in the noise shaping quantization calculation unit 24 according to the second embodiment, the compensation value calculation unit 34 and the compensation unit 35 are configured as a differentiator (1-z －１ In this case, the output v(z) of the noise shaping quantization calculation unit 24 is expressed by the following equation (10).

[0095]

[0096] This is the filter F ω (z) parameters A and B are set to A=1 and B=0, and F ω (z) = z －１ In other words, this is the filter F ω Time constant τ of (z) ｃ The sampling period τ ｓ Therefore, the noise shaping quantization calculation unit 24 is a dynamic quantizer as a whole, similar to the first embodiment. Also, the noise shaping quantization calculation unit 24 functions as a high-pass filter as a whole, similar to the first embodiment.

[0097] As mentioned above, the filter F ω (z) to F ω (z) = z －１ and differentiating the quantization noise ω(z) is called a ΔΣ modulator (delta-sigma modulator). That is, the noise shaping quantization calculation unit 24 can be configured by a ΔΣ modulator.

[0098] In this way, even when the noise shaping quantization calculation unit 24 is configured using a ΔΣ modulator, vibrations of the electric vehicle 100 are suppressed more effectively, similar to the first embodiment.

[0099] FIG. 12 shows the final torque command value T ｆｉｎ ＊ 12 is a graph showing the error ΔT of the final torque command value T ｆｉｎ ＊ The error ΔT and its amplitude spectrum are shown side by side.

[0100] That is, FIG. 12A shows the final torque command value T ｆｉｎ ＊ 12B is a graph showing the change over time of the error ΔT in the first comparative example. ｆｉｎ 12C is a graph showing the amplitude spectrum of the final torque command value T ｆｉｎ ＊ 12D is a graph showing the change over time of the error ΔT in the second embodiment. ｆｉｎ 10 is a graph showing the amplitude spectrum of

[0101] As shown in FIGS. 12A and 12C, when the noise shaping quantization calculation unit 24 is configured using a ΔΣ modulator, the final torque command value T ｆｉｎ ＊ 12B and 12D, in the first comparative example, the components of the error ΔT are concentrated in the low frequency band of 10 Hz or less, whereas in the second embodiment, the low frequency components of the error ΔT are reduced and the high frequency components of 10 Hz or more increase. These features are generally similar to those of the first embodiment.

[0102] FIG. 13 shows the longitudinal acceleration A Ｌ 13 is a graph showing the change over time of the longitudinal acceleration A of the first comparative example and the second embodiment, similarly to the first embodiment (FIG. 9). Ｌ The time changes of are shown side by side.

[0103] That is, FIG. 13A shows the longitudinal acceleration A occurring in the electric vehicle 100 in the first comparative example for the time range corresponding to FIG. 5. Ｌ 13B is an enlarged view of the area indicated by the dashed line in FIG. 13A. In FIGS. 13A and 13B, the longitudinal acceleration A Ｌ is shown by a solid line, and the ideal longitudinal acceleration A Ｌ The change in is shown by the dashed line.

[0104] FIG. 13C shows the longitudinal acceleration A of the electric vehicle 100 in the second embodiment for the time range corresponding to FIG. 5 . Ｌ 13(C) shows the change over time of the longitudinal acceleration A. In addition, FIG. 13(D) is an enlarged view of the range indicated by the dashed line in FIG. 13(C). The range indicated by the dashed line in FIG. 13(C) is the same as the range indicated by the dashed line in FIG. 13(A). In FIGS. 13(C) and 13(D), the longitudinal acceleration A Ｌ is shown by a solid line, and the ideal longitudinal acceleration A Ｌ The change in is shown by the dashed line.

[0105] As shown in FIGS. 13A to 13D, when the first comparative example is compared with the second embodiment, in the second embodiment, the time t ２ Hereafter, the upper limit torque command value T ｌｉｍ ＊ When the longitudinal acceleration A of the electric vehicle 100 gradually decreases, Ｌ In the second embodiment, vibrations that occur in the longitudinal acceleration A Ｌ These features are also generally similar to those of the first embodiment.

[0106] FIG. 14 shows the longitudinal acceleration A Ｌ Error ΔA Ｌ 14 is a graph showing the longitudinal acceleration A Ｌ Error ΔA Ｌ and its amplitude spectrum are shown side by side.

[0107] That is, FIG. 14A shows the longitudinal acceleration A Ｌ Error ΔAＬ 14B is a graph showing the change over time of the error ΔA Ｌ 14C is a graph showing the amplitude spectrum of the longitudinal acceleration A Ｌ Error ΔA Ｌ 14(D) is a graph showing the time change of the error ΔA Ｌ 10 is a graph showing the amplitude spectrum of

[0108] As shown in FIGS. 14A and 14C, when the noise shaping quantization calculation unit 24 is configured using a ΔΣ modulator, the longitudinal acceleration A Ｌ Error ΔA Ｌ This is also apparent from FIGS. 13(B) and 13(D).

[0109] Furthermore, as shown in FIGS. 14B and 14D, in the first comparative example, the error ΔA Ｌ The component of is widely distributed in the low frequency band of 10 Hz or less, whereas in the second embodiment, the error ΔA Ｌ is only slightly distributed in a frequency band of about 10 Hz, that is, near the resonant frequency of the vibrations generated in the electric vehicle 100. Therefore, even when the noise shaping quantization calculation unit 24 is configured using a ΔΣ modulator as in the second embodiment, the vibrations of the electric vehicle 100 are suppressed better than in the first comparative example.

[0110] [Third Embodiment] In the first and second embodiments, the noise-shaping quantization calculation unit 24 is configured by a dynamic quantizer, but this is not limiting. The noise-shaping quantization calculation unit 24 may be configured by a quantizer that performs noise shaping in another form.

[0111] 15 is a block diagram showing the configuration of the noise shaping quantization calculation unit 24 according to the third embodiment. As shown in FIG. 15, the noise shaping quantization calculation unit 24 according to the third embodiment is made up of a noise generator 40, a compensation unit 41, and a uniform quantizer 32.

[0112] The noise generator 40 generates a noise signal η (dither signal). The noise signal η is a random signal (so-called random noise) that follows a predetermined probability distribution. For example, the noise signal η is uniformly distributed noise, normally distributed noise, or triangularly distributed noise. In this embodiment, the noise signal η generated by the noise generator 40 is uniformly distributed noise. The magnitude of the noise signal η is determined adaptively depending on the magnitude of the quantization noise ω(z) or the quantization width d.

[0113] The compensation unit 41 adds a noise signal η to the input u(z) of the uniform quantizer 32. Therefore, the output v(z) of the noise shaping quantization calculation unit 24 of the third embodiment is expressed by the following equation (11).

[0114]

[0115] As described above, when the noise signal η is added to the input u(z), the quantization noise ω(z) superimposed on the output v(z) becomes somewhat uniform across frequency. Therefore, by adding the noise signal η to the input u(z), the quantization noise ω(z), which is biased toward the low-frequency band, essentially shifts the low-frequency components toward the high-frequency side. A quantizer that adds the noise signal η to the input u(z) in this way is called a random dither quantizer. In other words, the noise-shaping quantization calculation unit 24 can be configured using a random dither quantizer. Furthermore, when the noise-shaping quantization calculation unit 24 is configured using a random dither quantizer, vibrations of the electric vehicle 100 are suppressed more effectively than in the past.

[0116] FIG. 16 shows the final torque command value T ｆｉｎ ＊ 12 is a graph showing the error ΔT of the final torque command value T ｆｉｎ ＊ The error ΔT and its amplitude spectrum are shown side by side.

[0117] That is, FIG. 16A shows the final torque command value T ｆｉｎ＊ 16B is a graph showing the change over time of the error ΔT in the first comparative example. ｆｉｎ 16C is a graph showing the amplitude spectrum of the final torque command value T ｆｉｎ ＊ 16D is a graph showing the change over time of the error ΔT in the third embodiment. ｆｉｎ 10 is a graph showing the amplitude spectrum of

[0118] As shown in FIGS. 16A and 16C, when the noise shaping quantization calculation unit 24 is configured by a random dither quantizer, the final torque command value T ｆｉｎ ＊ The increase in the error ΔT is small.

[0119] 16B and 16D, in the first comparative example, the components of the error ΔT are concentrated in the low frequency band of 10 Hz or less, whereas in the third embodiment, the low frequency components of the error ΔT decrease and the high frequency components of 10 Hz or more increase. As a result, in the third embodiment, the error ΔT becomes roughly uniform with respect to frequency.

[0120] FIG. 17 shows the longitudinal acceleration A Ｌ 13 is a graph showing the change over time of the longitudinal acceleration A of the first comparative example and the third embodiment, similarly to the first embodiment (FIG. 9) and the second embodiment (FIG. 13). Ｌ The time changes of are shown side by side.

[0121] 17A shows the longitudinal acceleration A occurring in the electric vehicle 100 in the first comparative example for the time range corresponding to FIG. Ｌ 17(A) and 17(B) show the longitudinal acceleration A of the first comparative example. Ｌ is shown by a solid line, and the ideal longitudinal acceleration A Ｌ The change in is shown by the dashed line.

[0122] FIG. 17C shows the longitudinal acceleration A of the electric vehicle 100 in the third embodiment for the time range corresponding to FIG. 5 . Ｌ17(C) and 17(D) show the time change of the longitudinal acceleration A of this embodiment. Also, FIG. 17(D) is an enlarged view of the range indicated by the dashed line in FIG. 17(C). The range indicated by the dashed line in FIG. 17(C) is the same as the range indicated by the dashed line in FIG. 17(A). In FIGS. 17(C) and 17(D), the longitudinal acceleration A of this embodiment is Ｌ is shown by a solid line, and the ideal longitudinal acceleration A Ｌ The change in is shown by the dashed line.

[0123] As shown in FIGS. 17A to 17D, when the first comparative example is compared with the third embodiment, in the third embodiment, the time t ２ Hereafter, the upper limit torque command value T ｌｉｍ ＊ When the longitudinal acceleration A of the electric vehicle 100 gradually decreases, Ｌ Furthermore, in the third embodiment, the longitudinal acceleration A Ｌ These features are generally similar to those of the first and second embodiments.

[0124] FIG. 18 shows the longitudinal acceleration A Ｌ Error ΔA Ｌ 18 is a graph showing the longitudinal acceleration A and its amplitude spectrum for the first comparative example and the third embodiment, similarly to the first embodiment (FIG. 10) and the second embodiment (FIG. 14). Ｌ Error ΔA Ｌ and its amplitude spectrum are shown side by side.

[0125] That is, FIG. 18A shows the longitudinal acceleration A Ｌ Error ΔA Ｌ 18B is a graph showing the time change of the error ΔA Ｌ 18C is a graph showing the amplitude spectrum of the longitudinal acceleration A Ｌ Error ΔA Ｌ 18(D) is a graph showing the time change of the error ΔA Ｌ 10 is a graph showing the amplitude spectrum of

[0126] As shown in FIGS. 18A and 18C, when the noise shaping quantization calculation unit 24 is configured by a random dither quantizer, the longitudinal acceleration A Ｌ Error ΔA Ｌ However, as shown in FIGS. 18B and 18D, in the third embodiment, the error ΔA Ｌ In other words, the components near the resonant frequency of the vibrations occurring in the electric vehicle 100 are reduced. Therefore, even when the noise shaping quantization calculation unit 24 is configured using a random dither quantizer as in the third embodiment, the vibrations of the electric vehicle 100 are suppressed better than in the first comparative example.

[0127] [Fourth Embodiment] In the first and second embodiments, the noise-shaping quantization operation unit 24 is configured by a dynamic quantizer that compensates for the quantization noise ω(z) generated by the uniform quantizer 32, but this is not limiting. The noise-shaping quantization operation unit 24 can be configured to further compensate for the quantization noise ω(z) generated by the dynamic quantizer. In other words, the noise-shaping quantization operation unit 24 can be configured by a high-order (second-order or higher) dynamic quantizer.

[0128] 19 is a block diagram showing the configuration of the noise shaping quantization calculation unit 24 according to the fourth embodiment. As shown in FIG. 19, the noise shaping quantization calculation unit 24 according to the fourth embodiment is configured by a second-order dynamic quantizer 50.

[0129] The second-order dynamic quantizer 50 is a dynamic quantizer that estimates the quantization noise ω(z) generated by the first-order dynamic quantizer 51, calculates a compensation value based on the quantization noise ω(z), and uses this compensation value to compensate the output v(z) of the noise-shaping quantization calculation unit 24. Specifically, the second-order dynamic quantizer 50 is composed of the first-order dynamic quantizer 51, a quantization noise estimation unit 52, a compensation value calculation unit 53, and a compensation unit 54.

[0130] The primary dynamic quantizer 51 is composed of a uniform quantizer 32, a quantization noise estimation unit 33, a compensation value calculation unit 34, and a compensation unit 35. In other words, the primary dynamic quantizer 51 is the dynamic quantizer of the first or second embodiment, and is configured to compensate for the quantization noise ω(z) generated by the uniform quantizer 32.

[0131] The quantization noise estimation unit 52 estimates the quantization noise ω(z) caused by the primary dynamic quantizer 51 based on the input and output of the primary dynamic quantizer 51. Specifically, the quantization noise estimation unit 52 calculates the quantization noise ω(z) caused by the primary dynamic quantizer 51 by calculating the difference between the input and output of the primary dynamic quantizer 51.

[0132] The compensation value calculation unit 53 calculates a compensation value for the input u(z) based on the quantization noise ω(z). In this embodiment, the compensation value calculation unit 53 calculates a compensation value for the input u(z) based on the quantization noise ω(z). ω (z). Therefore, the compensation value is F ω (z)ω(z).

[0133] The compensation unit 54 compensates the output v(z) of the noise-shaping quantization operation unit 24 by subtracting a compensation value from the input u(z). Therefore, the output v(z) of the noise-shaping quantization operation unit 24 can be expressed as a whole by the following equation (12) using the order L of the dynamic quantizer. In the example shown in FIG. 19, L=2.

[0134]

[0135] Although FIG. 19 shows a second-order dynamic quantizer, a third-order or higher-order dynamic quantizer can also be configured in a similar manner. That is, if the dynamic quantizer that shapes the quantization noise ω(z) of the uniform quantizer 32 is a first-order dynamic quantizer, the noise-shaping quantization operation unit 24 can be configured with a second-order dynamic quantizer that further shapes the quantization noise ω(z) of this first-order dynamic quantizer. The noise-shaping quantization operation unit 24 can also be configured with a third-order dynamic quantizer that shapes the quantization noise ω(z) of the second-order dynamic quantizer. More generally, the noise-shaping quantization operation unit 24 can be configured with an L-th order dynamic quantizer that shapes the quantization noise ω(z) of an L-1-th order dynamic quantizer. In other words, the noise-shaping quantization operation unit 24 can be configured with a high-order dynamic quantizer in which dynamic quantizers are multiplexed.

[0136] 19, the compensation value calculation units 34 and 53 are the same as the filter F ω (z), but either one or both of these can be replaced with F ω (z) = z －１ The same applies when the noise shaping quantization calculation unit 24 is configured with a third-order or higher-order dynamic quantizer.

[0137] As described above, when the noise-shaping quantization calculation unit 24 is configured using a high-order dynamic quantizer, the low-frequency components of the quantization noise ω(z) are shifted to the high-frequency side with greater precision. In other words, the high-order dynamic quantizer further reduces the low-frequency components of the quantization noise ω(z). Therefore, vibrations of the electric vehicle 100 are more effectively suppressed than in the first or second embodiment.

[0138] [Modifications] In the first to third embodiments, as a typical example, the upper limit torque command value T ｌｉｍ ＊ Since the quantization noise ω(z) causes vibration in the electric vehicle 100, the upper limit torque command value T ｌｉｍ ＊ (Continuous upper limit torque command value T ｌｉｍ－ｃｏｎｔ ＊) is quantized by the noise shaping quantization calculation unit 24, but this is not limited to this. ｒｅｑ ＊ When transmitting and receiving the required torque command value T ｒｅｑ ＊ (Quantized required torque command value T ｒｅｑ－ｑｕａｎｔ ＊ ) may increase vibrations of the electric vehicle 100. In this case, the quantization calculation unit 23 can be configured to perform noise shaping in the same way as the noise shaping quantization calculation unit 24.

[0139] That is, the upper limit torque command value T ｌｉｍ ＊ and the required torque command value T ｒｅｑ ＊ When transmitting and receiving these, the upper limit torque command value T ｌｉｍ ＊ , required torque command value T ｒｅｑ ＊ , or both of these, it is preferable to perform noise shaping.

[0140] In addition, the required torque command value T ｒｅｑ ＊ and the upper limit torque command value T ｌｉｍ ＊ In addition to the above, there is a basic command value involved in vibration suppression control, and if the quantization noise ω(z) of the basic command value contributes to vibration of the electric vehicle 100, it is preferable to perform noise shaping when quantizing the basic command value.

[0141] As described above, the control method for an electric vehicle according to the first to third embodiments and the modified examples is a control method for an electric vehicle 100 that controls the operation of the rotating electric machine 11 using the first electronic control unit (13) and the second electronic control unit (15) that communicates with the first electronic control unit (13) via a communication network (CAN). In this control method, the first electronic control unit (13) controls the operation of the rotating electric machine 11 using a basic command value (T ｌｉｍ－ｃｏｎｔ ＊) is calculated, and the basic command value (T ｌｉｍ－ｃｏｎｔ ＊ ) is quantized, and the quantized basic command value (T ｌｉｍ－ｃｏｎｔ ＊ ) is the quantization basic command value (T ｌｉｍ－ｑｕａｎｔ ＊ ) is transmitted to the second electronic control unit (15) via a communication network (CAN). In addition, the second electronic control unit (15) transmits the quantized basic command value (T ｌｉｍ－ｑｕａｎｔ ＊ ) from the first electronic control unit (13), and the quantized basic command value (T ｌｉｍ－ｑｕａｎｔ ＊ ) to obtain a correction command value (T ｄａｍ ＊ ) is calculated, and the correction command value (T ｄａｍ ＊ ) and controls the operation of the rotating electrical machine 11 based on the basic command value (T ｌｉｍ－ｃｏｎｔ ＊ When quantizing ω(z), the quantization noise ω(z) is shaped so that predetermined frequency components including the resonance frequency of the vibration of the electric vehicle 100 move to the high frequency side.

[0142] In this way, in an electric vehicle 100 in which a basic command value used to calculate a vibration-damping correction command value is transmitted and received between a plurality of ECUs via a CAN, if the quantization noise ω(z) is shaped when the basic command value is quantized so that a predetermined frequency component including the resonance frequency of the vibration of the electric vehicle 100 moves to the higher frequency side, the vibration of the electric vehicle 100 can be more effectively suppressed. Specifically, the vibration of the electric vehicle 100 is more effectively suppressed than when the basic command value is simply quantized by the uniform quantizer 32 (first comparative example). Furthermore, since there is no delay as occurs when the quantized basic command value is processed by a low-pass filter on the receiving side (second ECU 15), torque limitation and the like are delayed, and the output power of the battery 10 does not reach the limited power P ｌｉｍ It is also possible to prevent the above from being exceeded.

[0143] In the control method for an electric vehicle according to the first, second and modified embodiments, the first electronic control unit (13) calculates a basic command value (T ｌｉｍ－ｃｏｎｔ ＊ ) and the quantization basic command value (T ｌｉｍ－ｃｏｎｔ ＊ ) and calculates the quantization noise ω(z) by the difference between the quantization noise ω(z) and the compensation value (F ω (z)ω(z)) and uses the compensation value to shape the quantization noise ω(z).

[0144] In this way, when the basic command value is quantized by the dynamic quantizer, vibrations of the electric vehicle 100 are particularly easily suppressed.

[0145] In the control method for an electric vehicle according to the first, second and modified embodiments, a filter F for extracting a predetermined frequency component is provided in the first electronic control unit (13). ω (z) to obtain the compensation value (F ω (z)ω(z)

[0146] Thus, the filter F ω If the compensation value is calculated using (z), it is possible to accurately extract the predetermined frequency component including the resonance frequency of the vibration of the electric vehicle 100. ω (z) does not substantially affect the input u(z) but only the quantization noise ω(z), so there is no delay in the output v(z). ω If the compensation value is calculated using (z), it is possible to suppress vibrations of the electric vehicle 100 in particular without causing delay.

[0147] In the control method for an electric vehicle according to the first embodiment and the modified example, the filter F ω (z) is composed of a low-pass filter.

[0148] Thus, the filter F ω (z) is the low-pass filter (F ω (z) = Az －１ / (1-Bz －１ )), as shown in FIG. 10 and other figures, vibrations of the electric vehicle 100 are particularly likely to be suppressed.

[0149] In the control method for an electric vehicle according to the second embodiment and the modification, a low-pass filter (F ω (z)) time constant τ ｃ is the sampling period τ ｓ is equal to.

[0150] Thus, τ ｃ = τ ｓ By setting ω (z) = z －１ Even when the basic command value is quantized using a ΔΣ modulator having the above configuration, vibrations of the electric vehicle 100 are likely to be suppressed, as shown in FIG. 14 and other figures.

[0151] In the control method for an electric vehicle according to the fourth embodiment, a dynamic quantizer of second or higher order that further shapes the quantization noise ω(z) generated by the dynamic quantizer is used to calculate the basic command value (T ｌｉｍ－ｃｏｎｔ ＊ ) is quantized.

[0152] In this way, when the basic command value is quantized using a high-order dynamic quantizer, the predetermined frequency components including the resonance frequency of the vibration of the electric vehicle 100 are shifted to the high frequency side with higher accuracy, and the low frequency components of the quantization noise ω(z) are further reduced. Therefore, the vibration of the electric vehicle 100 is particularly likely to be suppressed.

[0153] In the control method for an electric vehicle according to the third embodiment and the modified example, the first electronic control unit (13) generates a noise signal η and calculates a basic command value (T ｌｉｍ－ｃｏｎｔ ＊ ) to shape the quantization noise ω(z).

[0154] In this way, even when the basic command value is quantized by the random dither quantizer, the predetermined frequency component including the resonance frequency of the vibration of the electric vehicle 100 is substantially shifted to the higher frequency side. Therefore, the vibration of the electric vehicle 100 is more easily suppressed.

[0155] In the control method for an electric vehicle according to the first to third embodiments and the modified example, a plurality of basic command values ​​(T ｒｅｑ－ｃｏｎｔ ＊ , T ｌｉｍ－ｃｏｎｔ ＊) based on the correction command value (T ｄａｍ ＊ ), at least the basic command value (T ｌｉｍ－ｃｏｎｔ ＊ When quantizing ω(z), the quantization noise ω(z) is shaped.

[0156] In this way, if the quantization width d is set wide due to restrictions on the amount of communication in CAN communication and the quantization noise ω(z) of the roughly quantized basic command value is shaped, vibrations of the electric vehicle 100 can be easily suppressed.

[0157] In the control methods for electric vehicles according to the first to third embodiments and the modified examples, the corrected command value (T ｄａｍ ＊ ) is used as the basic command value for calculating the required torque command value (T ｒｅｑ－ｃｏｎｔ ＊ ) and the upper limit torque command value (T ｌｉｍ－ｃｏｎｔ ＊ ) and the required torque command value (T ｒｅｑ－ｃｏｎｔ ＊ ) quantization width (d ｒｅｑ ) than the upper limit torque command value (T ｌｉｍ－ｃｏｎｔ ＊ ) quantization width (d ｌｉｍ ) is wide, at least the upper limit torque command value (T ｒｅｑ－ｃｏｎｔ ＊ ) is quantized, the quantization noise ω(z) is shaped.

[0158] In this way, the required torque command value T ｒｅｑ ＊ and the upper limit torque command value T ｌｉｍ ＊ When calculating the upper limit torque command value T ｌｉｍ ＊ It is preferable to shape the quantization noise ω(z) of the upper limit torque command value T ｌｉｍ ＊ quantization width d (d ｌｉｍ ) is set widely, the upper limit torque command value T ｌｉｍ ＊ This is because the quantization noise ω(z) often causes vibrations in the electric vehicle 100.

[0159] The control device for an electric vehicle according to the first to third embodiments and the modified example is a control device (controller) for an electric vehicle 100 that controls the operation of a rotating electric machine 11 using a first electronic control unit (13) and a second electronic control unit (15) that communicates with the first electronic control unit (13) via a communication network (CAN). In this control device (controller), the first electronic control unit (13) controls a basic command value (T ｌｉｍ－ｃｏｎｔ ＊ ) is calculated, and the basic command value (T ｌｉｍ－ｃｏｎｔ ＊ ) is quantized, and the quantized basic command value (T ｌｉｍ－ｃｏｎｔ ＊ ) is the quantization basic command value (T ｌｉｍ－ｑｕａｎｔ ＊ ) to the second electronic control unit (15) via a communication network (CAN). The second electronic control unit (15) receives the quantized basic command value (T ｌｉｍ－ｑｕａｎｔ ＊ ) and receives the quantization basic command value (T ｌｉｍ－ｑｕａｎｔ ＊ ) to obtain a correction command value (T ｄａｍ ＊ ) and calculate the correction command value (T ｄａｍ ＊ ) and controls the operation of the rotating electrical machine 11 based on the basic command value (T ｌｉｍ－ｃｏｎｔ ＊ When quantizing ω(z), the quantization noise ω(z) is shaped so that predetermined frequency components including the resonance frequency of the vibration of the electric vehicle 100 move to the high frequency side.

[0160] In this way, in an electric vehicle 100 in which a basic command value used to calculate a vibration-damping correction command value is transmitted and received between a plurality of ECUs via a CAN, if the quantization noise ω(z) is shaped when the basic command value is quantized so that a predetermined frequency component including the resonance frequency of the vibration of the electric vehicle 100 moves to the higher frequency side, the vibration of the electric vehicle 100 can be more effectively suppressed. Specifically, the vibration of the electric vehicle 100 is more effectively suppressed than when the basic command value is simply quantized by the uniform quantizer 32 (first comparative example). Furthermore, since there is no delay as occurs when the quantized basic command value is processed by a low-pass filter on the receiving side (second ECU 15), torque limitation and the like are delayed, and the output power of the battery 10 does not reach the limited power P ｌｉｍ It is also possible to prevent the above from being exceeded.

[0161] The above describes embodiments and modifications of the present invention, but the configurations described in the above embodiments and modifications merely illustrate some of the application examples of the present invention and are not intended to limit the technical scope of the present invention.

Claims

1. A control method for an electric vehicle, which controls the operation of a rotating electric machine using a first electronic control unit and a second electronic control unit that communicates with the first electronic control unit via a communication network, wherein the first electronic control unit: calculates basic command values ​​that determine the operating state of the rotating electric machine; quantizes the basic command values ​​with a quantization width that corresponds to the constraints of the communication volume in the communication network; and transmits quantized basic command values, which are the quantized basic command values, to the second electronic control unit via the communication network; the second electronic control unit: receives the quantized basic command values ​​from the first electronic control unit; uses the quantized basic command values ​​to calculate corrected command values ​​that suppress vibrations of the electric vehicle; and controls the operation of the rotating electric machine based on the corrected command values; and when the basic command values ​​are quantized in the first electronic control unit, quantization noise is shaped so that predetermined frequency components including a resonance frequency of the vibrations are shifted to higher frequencies.

2. A control method for an electric vehicle according to claim 1, wherein the first electronic control unit calculates the quantization noise from the difference between the basic command value and the quantized basic command value, calculates a compensation value based on the quantization noise, and shapes the quantization noise using the compensation value.

3. A control method for an electric vehicle according to claim 2, wherein the first electronic control unit calculates the compensation value from the quantization noise using a filter that extracts the predetermined frequency component.

4. A method for controlling an electric vehicle according to claim 3, wherein the filter is configured by a low-pass filter.

5. A method for controlling an electric vehicle according to claim 4, wherein the time constant of the low-pass filter is equal to the sampling period.

6. A method for controlling an electric vehicle according to claim 2, wherein the basic command value is quantized by a second-order or higher dynamic quantizer that further shapes the quantization noise generated by the dynamic quantizer.

7. A control method for an electric vehicle according to claim 1, wherein the first electronic control unit generates a noise signal, and quantizes the basic command value to which the noise signal has been added, thereby shaping the quantized noise.

8. A control method for an electric vehicle as claimed in claim 1, wherein, when the corrected command value is calculated based on a plurality of basic command values ​​having different quantization widths, the quantization noise is shaped at least when the basic command value having a relatively wide quantization width is quantized.

9. A control method for an electric vehicle as claimed in claim 8, wherein a required torque command value and an upper limit torque command value are calculated as the basic command values ​​used in calculating the corrected command value, and when the quantization width of the upper limit torque command value is wider than the quantization width of the required torque command value, the quantization noise is shaped at least when quantizing the upper limit torque command value.

10. A control device for an electric vehicle that controls the operation of a rotating electric machine using a first electronic control unit and a second electronic control unit that communicates with the first electronic control unit via a communication network, wherein the first electronic control unit is configured to: calculate basic command values ​​that determine the operating state of the rotating electric machine, quantize the basic command values ​​with a quantization width according to the constraints of the communication volume in the communication network, and transmit quantized basic command values, which are the quantized basic command values, to the second electronic control unit via the communication network, wherein the second electronic control unit is configured to: receive the quantized basic command values ​​from the first electronic control unit, use the quantized basic command values ​​to calculate corrected command values ​​that suppress vibrations of the electric vehicle, and control the operation of the rotating electric machine based on the corrected command values, and wherein the first electronic control unit is configured to shape quantization noise when quantizing the basic command values ​​so that predetermined frequency components including a resonance frequency of the vibrations move toward higher frequencies.

Images