Electric vehicle control device

Abstract

To start the regulation of a power factor of electric power between an inverter and a motor without delay in an electric vehicle.SOLUTION: An execution device executes: processing of determining whether the value of an operation variable which varies in response to an occupant's operation reaches a determination value (Step S110); processing of calculating the value of an output variable by inputting, into mapping, the values of input variables when the value of the operation variable reaches the determination value (Step S140); processing of determining whether a motor is brought into a high-torque state, on the basis of the value of the output variable (Step S150); and processing of, when determining that the motor is brought into the high-torque state, outputting a start signal for starting the regulation of a power factor of electric power between an inverter and the motor (Step S170). The input variables are variables indicating travel information on an electric vehicle when the value of the operation variable reaches the determination value. The output variable is a variable indicating whether the motor is brought into the high-torque state when a fixed time elapses after the operation variable reaches the determination value.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to a control device for an electric vehicle. [Background technology]

[0002] The electric vehicle disclosed in Patent Document 1 has a battery, an inverter, a motor, a power factor adjustment circuit, and a control device. The battery is connected to the motor via the inverter. The inverter converts DC to AC between the battery and the motor. The motor receives power from the battery via the inverter. This drives the motor. The power factor adjustment circuit is connected in parallel to the coil of the motor. The power factor adjustment circuit includes a thyristor and a capacitor. The capacitance of the capacitor changes depending on the length of time that the thyristor is turned on. The control device adjusts the capacitance of the capacitor through control of the thyristor. In this way, the control device adjusts the power factor of the power supplied to the motor. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2012-147572 A Summary of the Invention [Problem to be solved by the invention]

[0004] In an electric vehicle such as that described in Patent Document 1, when the torque of the motor is high, the amount of power supplied to the motor increases. From the viewpoint of reducing the power consumption of the battery, it is desirable to minimize reactive power under such a situation where the amount of power supplied to the motor is large. Therefore, it is conceivable to adjust the power factor of the power supplied to the motor so that it approaches 1 when the torque of the motor increases.

[0005] Here, in order to start the power factor adjustment, pre-processing is required, such as judging whether or not to start the power factor adjustment, and outputting a signal to start the power factor adjustment after the judgment. Here, let us assume that the above judgment is made by referring to the motor current value and judging that the motor torque has actually increased. In this case, the pre-processing is performed after the motor torque has actually increased. Therefore, a time lag occurs between the time when the motor torque increases and the power factor of the power is actually adjusted, which is the time required to execute the pre-processing. Therefore, during the time lag, a state in which the reactive power is high continues. [Means for solving the problem]

[0006] A control device for an electric vehicle for solving the above problems is applied to an electric vehicle having a battery, a motor that receives power from the battery, an inverter that performs DC / AC conversion between the battery and the motor, and a power factor adjustment circuit that adjusts the power factor of the power between the inverter and the motor, and the control device has a storage device and an execution device, the storage device pre-stores mapping data that specifies a mapping that has been learned by machine learning and that outputs a value of an output variable when values ​​of a plurality of input variables are input, the plurality of input variables are variables that indicate driving information of the electric vehicle when a value of an operation variable, the value of which changes in response to an operation of the electric vehicle by a occupant, reaches a predetermined judgment value, and the output variable is a value that indicates driving information of the electric vehicle when a certain time period after which the operation variable reaches the judgment value is reached, the operation variable is a variable that indicates whether or not the motor will later enter a high-torque state, and the execution device executes a first determination process of determining whether or not the value of the operation variable has reached the determination value; an acquisition process of acquiring values ​​of the multiple input variables when it is determined in the first determination process that the value of the operation variable has reached the determination value; a calculation process of calculating the value of the output variable by inputting the values ​​of the multiple input variables acquired by the acquisition process to the mapping; a second determination process of determining whether or not the motor will enter a high-torque state based on the value of the output variable calculated in the calculation process; and a signal output process of outputting a start signal to the power factor adjustment circuit to start adjusting the power factor so that the power factor approaches 1 when it is determined in the second determination process that the motor will enter a high-torque state.

[0007] In the above configuration, by using the mapping, the value of the output variable can be calculated based on not only the information that the value of the manipulated variable has reached the judgment value, but also the driving information of the electric vehicle when the value of the manipulated variable has reached the judgment value. In this case, if the mapping is properly learned, it is possible to accurately predict that the motor will enter a high torque state a certain time after the value of the manipulated variable has reached the judgment value or more. In other words, the future state of the motor can be predicted based on the driving information at the time when the value of the manipulated variable has reached the judgment value. In this way, since it is possible to know that the motor will enter a high torque state at a stage before it enters a high torque state, a start signal can be output without delay before the motor enters a high torque state. Therefore, the power factor can be controlled to a value close to 1 from the timing when the motor enters a high torque state. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a schematic diagram of a vehicle. [Diagram 2] FIG. 2 is a flowchart showing a specific processing procedure of the power factor adjustment process. [Diagram 3] FIG. 3 is a diagram showing a schematic diagram of a manner in which the server collects data. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, an embodiment of a control device for an electrically-equipped vehicle will be described with reference to the drawings. <Outline of electric vehicle configuration> As shown in FIG. 1, an electric vehicle (hereinafter simply referred to as a vehicle) 10 has a motor generator 20, an inverter 15, and a battery 11. The motor generator 20 is a drive source for the vehicle 10. The motor generator 20 functions as both an electric motor and a generator. The motor generator 20 has a rotor, a stator, a U-phase coil 20A, a V-phase coil 20B, and a W-phase coil 20C. Note that the stator and rotor are not shown in FIG. 1. Each of the coils 20A, 20B, and 20C is wound around the stator. The rotor rotates in response to the passage of current through each of the coils 20A, 20B, and 20C.

[0010] Each of the coils 20A, 20B, 20C of the motor generator 20 is connected to an inverter 15. The inverter 15 is connected to a positive terminal and a negative terminal of a battery 11. The battery 11 is a secondary battery. The battery 11 supplies power to the motor generator 20 via the inverter 15. The battery 11 also receives power from the motor generator 20 via the inverter 15. The battery 11 stores the supplied power.

[0011] The inverter 15 converts DC to AC between the motor generator 20 and the battery 11. Although not shown, the inverter 15 has an arm circuit for each coil of the motor generator 20. The arm circuit has two transistors connected in series and a freewheeling diode connected in parallel to each transistor. The transistors are npn-type switching elements that are switched on and off in response to an input of a control signal. The period of the AC current flowing from the inverter 15 to the motor generator 20 changes in accordance with the period of switching the transistors on and off.

[0012] The vehicle 10 has a power factor adjustment circuit 30. The power factor adjustment circuit 30 has a first circuit 30A, a second circuit 30B, and a third circuit 30C. The first circuit 30A has a first thyristor 31A, a second thyristor 32A, and a capacitor 33A. The first thyristor 31A is a switching element in which semiconductors are stacked in the order of p-type, n-type, p-type, and n-type. The first thyristor 31A has three terminals, namely, a cathode, an anode, and a gate. The second thyristor 32A is a switching element similar to the first thyristor 31A. The anode of the first thyristor 31A is connected to the U-phase coil 20A. The cathode of the first thyristor 31A is connected to a first terminal of the capacitor 33A. The second thyristor 32A is connected in parallel with the first thyristor 31A. The second thyristor 32A is oriented in the opposite direction to the first thyristor 31A. That is, the anode of the second thyristor 32A is connected to the cathode of the first thyristor 31A. The cathode of the second thyristor 32A is connected to the anode of the first thyristor 31A. The connection destination of the second terminal of the capacitor 33A will be described later.

[0013] The second circuit 30B has a first thyristor 31B, a second thyristor 32B, and a capacitor 33B, similar to the first circuit 30A. The configuration of the second circuit 30B is the same as that of the first circuit 30A, except for the following point. That is, the anode of the first thyristor 31B is connected to the V-phase coil 20B. The description of the other configuration of the second circuit 30B is omitted.

[0014] The third circuit 30C has a first thyristor 31C, a second thyristor 32C, and a capacitor 33C, similar to the first circuit 30A. The configuration of the third circuit 30C is the same as that of the first circuit 30A, except for the following point. That is, the anode of the first thyristor 31C is connected to the W-phase coil 20C. The description of the other configuration of the third circuit 30C is omitted.

[0015] The second terminal of the capacitor 33A in the first circuit 30A, the second terminal of the capacitor 33B in the second circuit 30B, and the second terminal of the capacitor 33C in the third circuit 30C are connected to each other.

[0016] The first circuit 30A functions as a variable capacitor capable of adjusting the capacitance. The reason is as follows. As a premise, since the current flowing through the first circuit 30A is an alternating current, the value of the current flowing through the first circuit 30A alternates between positive and negative. The first thyristor 31A has a characteristic of being turned on only when a positive current flows through the first circuit 30A. That is, it is assumed that a positive current flows through the first thyristor 31A. At this time, when a control signal is input to the gate, the first thyristor 31A turns on. Once the first thyristor 31A turns on, the first thyristor 31A remains on as long as the value of the current flowing through the first thyristor 31A is positive, even if the control signal to the gate is no longer present. Then, when the value of the current flowing through the first thyristor 31A becomes zero according to the phase of the current, the first thyristor 31A turns off automatically. Due to the characteristics of the first thyristor 31A, the period during which the first thyristor 31A is on varies depending on the timing of input of a control signal to the gate in one cycle of the AC current. Hereinafter, the timing of input of a control signal to the gate is referred to as the firing phase for the thyristor. Contrary to the first thyristor 31A, the second thyristor 32A has a characteristic of being on during a period during which a negative current flows through the first circuit 30A. Here, the amount of charge stored in the capacitor 33A varies depending on the period during which each of the thyristors 31A and 32A is on. Therefore, the charge stored in the capacitor 33A varies depending on the firing phase for each of the thyristors 31A and 32A. In other words, the capacitance of the capacitor 33A can be adjusted by adjusting the firing phase for each of the thyristors 31A and 32A. Like the first circuit 30A, the second circuit 30B and the third circuit 30C also function as variable capacitors.

[0017] When viewed from the inverter 15, the first circuit 30A is equivalent to being connected in parallel to the U-phase coil 20A in the motor generator 20. In this case, the phase of the current flowing through the U-phase coil 20A relative to the voltage applied to the coil 20A changes depending on the capacitance of the first circuit 30A. This means that the power factor of the power between the inverter 15 and the motor generator 20 changes. In other words, the first circuit 30A can change the power factor, which is the ratio of active power to apparent power output from the inverter 15 to the motor generator 20. The above description using the U-phase as an example also applies to the V-phase and W-phase.

[0018] The vehicle 10 has an accelerator pedal 60. The accelerator pedal 60 is a foot pedal that is depressed by a vehicle occupant. The vehicle 10 also has a start switch 44. The start switch 44 is a switch for instructing system on / system off of the vehicle 10.

[0019] The vehicle 10 has a GPS receiver 40, a vehicle speed sensor 42, and an accelerator sensor 48 as information acquisition devices. The GPS receiver 40 receives a signal related to the current position coordinate G of the vehicle 10 from a GPS satellite. That is, the GPS receiver 40 receives a signal related to the current latitude and longitude coordinate values ​​of the vehicle 10. The vehicle speed sensor 42 detects the vehicle speed SP, which is the traveling speed of the vehicle 10. The accelerator sensor 48 detects the accelerator operation amount ACC, which is the operation amount of the accelerator pedal 60. The information acquisition devices repeatedly output the information that they detect and acquire to a control device 90, which will be described later.

[0020] <Overview of the control device> The vehicle 10 has a control device 90. The control device 90 may be configured as one or more processors that execute various processes according to a computer program (software). The control device 90 may be configured as one or more dedicated hardware circuits, such as an application specific integrated circuit (ASIC), that execute at least a part of the various processes, or a circuit including a combination thereof. The processor includes a CPU 91 and memories such as a RAM and a ROM 93. The memory stores program codes or instructions that are configured to cause the CPU 91 to execute processes. The memory, i.e., the computer-readable medium, includes any available medium that can be accessed by a general-purpose or dedicated computer. In this embodiment, the CPU 91 and the ROM 93 constitute an execution device.

[0021] The control device 90 has a communication device 94, a storage device 95, a real-time clock 96, and an internal bus 98. The communication device 94 is a device for communicating with the outside of the vehicle 10 via an external communication line network 100. The storage device 95 is an electrically rewritable non-volatile memory. The real-time clock 96 is a circuit that generates date and time information. The internal bus 98 enables the CPU 91, ROM 93, communication device 94, storage device 95, and real-time clock 96 to communicate with one another.

[0022] The control device 90 repeatedly receives information output by various information acquisition devices attached to the vehicle 10. The control device 90 also communicates with a weather service center 500 via an external communication network 100 by utilizing the functions of a communication device 94. The control device 90 then repeatedly receives weather information J distributed from the weather service center 500. The weather information J includes information on the weather around the location where the vehicle 10 is traveling. The control device 90 also receives a signal W in response to the operation of the start switch 44.

[0023] The CPU 91 executes programs stored in the ROM 93 to control various components of the vehicle 10. For example, the CPU 91 controls the motor generator 20. Specifically, the CPU 91 calculates a vehicle required output, which is a required value of an output necessary for the vehicle 10 to travel, based on the accelerator operation amount ACC and the vehicle speed SP. Then, the CPU 91 calculates a required torque for the motor generator 20 based on the vehicle required output. The CPU 91 controls the inverter 15 so that the torque of the motor generator 20 becomes the required torque.

[0024] The storage device 95 stores a history of the required torque calculated by the CPU 91 for a certain period of time, overwriting the oldest one with the newest one. Similarly, the storage device 95 stores a history of the information received from the information acquisition device and the meteorological service center 500 for a certain period of time. Therefore, the storage device 95 stores not only the latest value for each parameter, but also values ​​calculated or received at, for example, the previous timing.

[0025] <Outline of power factor adjustment process> The CPU 91 can execute a power factor adjustment process. Here, when the torque of the motor generator 20 is high, the power consumption of the battery 11 is higher than when the torque of the motor generator 20 is low. If the power consumption of the battery 11 is high, the reactive power increases accordingly. Therefore, in order to suppress the power consumption of the battery 11, it is desirable to suppress the reactive power as much as possible by making the power factor between the inverter 15 and the motor generator 20 close to 1 under a condition in which the power consumption is high, that is, under a condition in which the torque of the motor generator 20 is high. The power factor adjustment process is a process for adjusting the power factor between the inverter 15 and the motor generator 20 close to 1 when the motor generator 20 is in a high torque state. Note that the high torque state is a state in which the torque of the motor generator 20 is equal to or greater than a specified torque N. The specified torque N is determined, for example, by an experiment or a simulation, as a minimum torque value at which the output power required of the battery 11 according to the required torque is correspondingly large with respect to the capacity of the battery 11 mounted on the vehicle 10 and at which it is preferable to suppress the reactive power.

[0026] The storage device 95 pre-stores firing information as information necessary for bringing the power factor between the inverter 15 and the motor generator 20 closer to 1. The firing information is information about an adjustment firing phase for each thyristor in the power factor adjustment circuit 30. The adjustment firing phase is information about an adjustment firing phase for realizing a capacitance necessary for bringing the power factor of the power between the inverter 15 and the motor generator 20 closer to 1 when the motor generator 20 is in a high torque state. The adjustment firing phase is determined in advance, for example, by an experiment or a simulation.

[0027] The power factor adjustment process includes predicting in advance that the motor generator 20 will be in a high torque state. As information necessary for this prediction in advance, the storage device 95 stores mapping data in advance. The mapping data specifies a mapping D that outputs a value of an output variable by inputting values ​​of a plurality of input variables. This mapping D is a learned model that has been machine-learned before being implemented in the vehicle 10. The input variables are variables that indicate the driving information of the vehicle 10 when the accelerator operation amount ACC increases to a judgment value AK. The output variables are variables that indicate whether the motor generator 20 will be in a high torque state after a certain time T after the accelerator operation amount ACC increases to the judgment value AK. The judgment value AK and the certain time T will be described later. The accelerator operation amount ACC is an operation variable whose value changes depending on the depression of the accelerator pedal 60 by the occupant.

[0028] Specifically, the variables indicating the driving information of the vehicle 10 are the following variables. That is, each of these variables becomes an input variable to the mapping D. A location variable that indicates the location where the vehicle 10 is traveling. A time variable that indicates the time when the vehicle 10 is traveling. A seasonal variable that indicates the season in which the vehicle 10 is traveling. A weather variable that indicates the weather at the location where the vehicle 10 is traveling. The CPU 91 can execute a first determination process, an acquisition process, a calculation process, a second determination process, a pre-processing, a signal output process, and an adjustment continuation process as part of the power factor adjustment process. In the first determination process, the CPU 91 determines whether the accelerator operation amount ACC has increased to a determination value AK. In the acquisition process, the CPU 91 acquires the values ​​of a plurality of input variables when it is determined in the first determination process that the accelerator operation amount ACC has increased to the determination value AK. In the calculation process, the CPU 91 calculates the value of an output variable by inputting the values ​​of the plurality of input variables acquired in the acquisition process to a mapping D. In the second determination process, the CPU 91 determines whether the motor generator 20 will be in a high torque state based on the values ​​of the output variables calculated in the calculation process. In the pre-processing, the CPU 91 acquires ignition information when it is determined in the second determination process that the motor generator 20 will be in a high torque state. In the signal output process, the CPU 91 outputs a start signal to the power factor adjustment circuit 30 at an appropriate timing based on the ignition information acquired in the pre-processing. The start signal is a signal that instructs the start of power factor adjustment so that the power factor between the inverter 15 and the motor generator 20 approaches 1. The start signal is essentially an initial control signal that the CPU 91 outputs to the thyristor when adjusting the power factor. In the adjustment continuation process, the CPU 91 repeats output of the control signal to the thyristor to continue the adjustment of the power factor between the inverter 15 and the motor generator 20.

[0029] <Regarding fixed time and judgment value> As will be described in detail later, in the power factor adjustment process, when the CPU 91 determines in the first determination process that the accelerator operation amount ACC has increased to the determination value AK, it sequentially performs the acquisition process, the calculation process, the second determination process, the pre-processing, and the signal output process in this order. Then, when the CPU 91 outputs a start signal in the signal output process, the thyristor is switched on accordingly. The time required from the start of the acquisition process until the thyristor is switched on in response to the start signal is called the required time. The required time can be determined, for example, by an experiment or a simulation. The above-mentioned certain time T is set to be a time slightly longer than the required period.

[0030] The judgment value AK related to the accelerator operation amount ACC is a value determined based on the above-mentioned certain time T and the above-mentioned specified torque N that defines the high torque state. Here, the torque of the motor generator 20 increases according to the accelerator operation amount ACC. The above-mentioned judgment value AK is determined, for example, by an experiment or a simulation, as the accelerator operation amount ACC when going back by the above-mentioned certain time T from the timing when the torque of the motor generator 20 increased to the specified torque N. Specifically, the judgment value AK can be obtained as follows. A situation in which the motor generator 20 increases to the specified torque N is, for example, a situation in which the accelerator pedal 60 is quickly depressed to rapidly accelerate the vehicle 10 or to restore the vehicle speed SP on an uphill road. A relationship that can be generally regarded as a relationship between the time change of the accelerator operation amount ACC and the associated time change of the torque of the motor generator 20 for such a situation in which the accelerator pedal 60 is quickly depressed is created based on an experiment or a simulation. Then, based on the relationship, the judgment value AK is calculated backward from the timing when the torque of the motor generator 20 becomes the specified torque N. The storage device 95 stores the judgment value AK as information to be used in the first judgment process.

[0031] <Specific steps for power factor adjustment processing> The CPU 91 starts the power factor adjustment process when the start switch 44 is operated to turn on the system of the vehicle 10. The CPU 91 repeatedly performs the power factor adjustment process from when the start switch 44 is operated to turn on the system of the vehicle 10 until the start switch 44 is operated again to turn off the system of the vehicle 10. Note that when the system of the vehicle 10 is turned on, the CPU 91 does not output a control signal to the thyristor of the power factor adjustment circuit 30.

[0032] As shown in FIG. 2, when the CPU 91 starts the power factor adjustment process, it first performs the process of step S110. In step S110, the CPU 91 determines whether the accelerator operation amount ACC has increased to the judgment value AK. The CPU 91 makes the determination of step S110 based on the history of the accelerator operation amount ACC. Specifically, the CPU 91 acquires the latest accelerator operation amount ACC stored in the storage device 95, the accelerator operation amount ACC at the immediately previous timing, and the judgment value AK. The CPU 91 then compares the magnitudes of these acquired values. If the following first judgment item is satisfied, the CPU 91 determines that the accelerator operation amount ACC has increased to the judgment value AK, and if not, determines that the accelerator operation amount ACC has not increased to the judgment value AK. The first judgment item is that the accelerator operation amount ACC at the immediately previous timing is smaller than the judgment value AK, and the latest accelerator operation amount ACC is equal to or larger than the judgment value AK. If the first determination item is not satisfied (step S110: NO), the CPU 91 ends the series of steps of the power factor adjustment process. In this case, the CPU 91 executes the process of step S110 again.

[0033] On the other hand, if the first determination item is met (step S110: YES), the CPU 91 advances the process to step S120. The process of step S110 is the first determination process.

[0034] In step S120, the CPU 91 acquires the values ​​of each variable to be used in the subsequent processing. Specifically, the CPU 91 acquires the latest latitude Z1, the latest longitude Z2, the time identification value Z3, the month identification value Z4, and the weather identification value Z5. Below, the manner in which the CPU 91 acquires the values ​​of each of these variables will be described in order. Note that the values ​​of each variable acquired here are essentially the values ​​of the input variables to the mapping D.

[0035] The latest latitude Z1 is the latitude at which the vehicle 10 is located at the time when the CPU 91 performs the process of step S120. The CPU 91 acquires, as the latest latitude Z1, the coordinate value of the latest latitude stored in the storage device 95. The latest latitude Z1 is the above-mentioned location variable.

[0036] The latest longitude Z2 is the longitude where the vehicle 10 is located at the time when the CPU 91 performs the process of step S120. The CPU 91 acquires, as the latest longitude Z2, the coordinate value of the latest longitude stored in the storage device 95. The latest longitude Z2 is the above-mentioned point variable.

[0037] The time identification value Z3 represents the time of day in hourly units, and takes on values ​​from "1" to "24." That is, the time identification value Z3 is "1" for the period from midnight to 1:00 a.m., and the value increases every hour thereafter. The CPU 91 calculates the time identification value Z3 according to the time at the timing of performing the processing of step S120, based on the latest time information output by the real-time clock 96. The calculation of the time identification value Z3 by the CPU 91 corresponds to the CPU 91 acquiring the time identification value Z3. The time identification value Z3 is the above-mentioned time variable.

[0038] The month identification value Z4 represents each month of the year and takes on values ​​from "1" to "12." That is, the month identification value Z4 is increased each month, with January being "1." The CPU 91 calculates the month identification value Z4 according to the month at the time of performing the processing of step S120, based on the latest date information output by the real-time clock 96. The calculation of the month identification value Z4 by the CPU 91 corresponds to the CPU 91 acquiring the month identification value Z4. The month identification value Z4 is the above-mentioned seasonal variable.

[0039] The weather identification value Z5 is an identification value for identifying the type of weather. In this embodiment, there are four types of weather: "sunny", "cloudy", "rain", and "snow". The weather identification value Z5 is set to "1" for "sunny", "2" for "cloudy", "3" for "rain", and "4" for "snow". The CPU 91 handles the latest weather information J stored in the storage device 95 as the weather information J at the time of performing the processing of step S120. The CPU 91 calculates the weather identification value Z5 according to the type of weather around the vehicle 10, which is included in the weather information J. The calculation of the weather identification value Z5 by the CPU 91 corresponds to the acquisition of the weather identification value Z5 by the CPU 91. The weather identification value Z5 is the above-mentioned weather variable.

[0040] After acquiring the values ​​of the variables in the above manner, the CPU 91 advances the process to step S130. Note that the process of step S120 is an acquisition process. In step S130, as preprocessing for performing a calculation using mapping D, the CPU 91 assigns the values ​​of each variable acquired in step S120 to input variables x(1) to x(5) for mapping D. Specifically, the CPU 91 assigns the latest latitude Z1 to input variable x(1). The CPU 91 assigns the latest longitude Z2 to input variable x(2). The CPU 91 assigns the time identification value Z3 to input variable x(3). The CPU 91 assigns the month identification value Z4 to input variable x(4). The CPU 91 assigns the weather identification value Z5 to input variable x(5). After this, the CPU 91 advances the process to step S140.

[0041] In step S140, the CPU 91 inputs the values ​​of the input variables x(1) to x(5) into the mapping D defined by the mapping data stored in the storage device 95. As a result, the CPU 91 calculates the values ​​of the output variables Q(1) to Q(2). The output variable Q(1) is a first probability R1 which is the probability that the motor generator 20 will be in a high torque state after a certain time T after the accelerator operation amount ACC has increased to the judgment value AK. In other words, the output variable Q(1) quantifies, as a value ranging from "0" to "1", the degree of likelihood that the motor generator 20 will be in a high torque state after the time T. The output variable Q(2) is a second probability R2 which is the probability that the motor generator 20 will not be in a high torque state after a certain time T after the accelerator operation amount ACC has increased to the judgment value AK. In other words, the output variable Q(2) quantifies, as a value ranging from "0" to "1", the degree of likelihood that the motor generator 20 will not be in a high torque state after the time T. The first probability R1 and the second probability R2 are variables that indicate whether or not the motor generator 20 will enter a high torque state after a certain time T has elapsed since the accelerator operation amount ACC increased to the determination value AK.

[0042] In this embodiment, the mapping D is composed of a fully connected forward propagation type neural network with one intermediate layer and a softmax function that transforms the output of the neural network. The neural network includes an input side coefficient wFjk (j=0 to n, k=0 to 5) and an activation function h(x) as an input side nonlinear mapping that nonlinearly transforms each of the outputs of the input side linear mapping, which is a linear mapping defined by the input side coefficient wFjk. In this embodiment, the activation function h(x) is exemplified as a hyperbolic tangent "tanh(x)". In addition, the neural network includes an output side coefficient wSij (i=1 to 2, j=0 to n) and an activation function f(x) as an output side nonlinear mapping that nonlinearly transforms each of the outputs of the output side linear mapping, which is a linear mapping defined by the output side coefficient wSij. In this embodiment, the activation function f(x) is exemplified as a hyperbolic tangent "tanh(x)". Note that the value n indicates the dimension of the intermediate layer. Moreover, the input side coefficient wFj0 ​​is a bias parameter, and is the coefficient of the input variable x(0). The input variable x(0) is defined as "1." Moreover, the output side coefficient wSi0 is a bias parameter.

[0043] The softmax function is a function that normalizes the output of the neural network to make the sum of the output variable Q(1) and the output variable Q(2) "1." As a specific process of step S140, the CPU 91 first inputs the values ​​of input variables x(1) to x(5) to the neural network. Then, the CPU 91 calculates the values ​​of stochastic prototypes y(1) to y(2) which are the output of the neural network. The stochastic prototype y(1) is a parameter that has a positive correlation with the probability that the motor generator 20 will be in a high torque state after a certain time T after the accelerator operation amount ACC has increased to the judgment value AK. The stochastic prototype y(2) is a parameter that has a positive correlation with the probability that the motor generator 20 will not be in a high torque state after a certain time T after the accelerator operation amount ACC has increased to the judgment value AK. After calculating the values ​​of the stochastic prototypes y(1) to y(2), the CPU 91 inputs them to a softmax function to calculate the values ​​of the output variables Q(1) to Q(2). After this, the CPU 91 advances the process to step S150. The process of step S140 is a calculation process.

[0044] In step S150, the CPU 91 determines whether or not the motor generator 20 will be in a high torque state based on the first probability R1, which is the output variable Q(1). If the first probability R1 is equal to or less than a threshold, the CPU 91 determines that the motor generator 20 will not be in a high torque state. In this case (step S150: NO), the CPU 91 ends the series of steps of the power factor adjustment process and executes the process of step S110 again. The threshold is, for example, "0.5."

[0045] On the other hand, in step S150, when the first probability R1 is greater than the threshold value, the CPU 91 determines that the motor generator 20 will be in a high torque state. In this case (step S150: YES), the CPU 91 advances the process to step S160. The process of step S150 is a second determination process.

[0046] In step S160, the CPU 91 performs a pre-processing for starting the adjustment of the power factor. Specifically, the CPU 91 acquires the firing information stored in the storage device 95. As described above, the firing information is information on the firing phase for adjustment of the thyristor of the power factor adjustment circuit 30, that is, the firing phase required to bring the power factor closer to 1 when the motor generator 20 is in a high torque state. Upon acquiring the firing information, the CPU 91 generates an initial control signal for the thyristor. After this, the CPU 91 advances the process to step S170.

[0047] In step S170, the CPU 91 outputs the initial control signal generated in step S160. As described above, this control signal is a start signal that instructs the power factor adjustment circuit 30 to start adjusting the power factor. When outputting the start signal, the CPU 91 grasps the phase of the current flowing through the power factor adjustment circuit 30 based on the switching status of the inverter 15. Then, the CPU 91 outputs the start signal at an appropriate timing based on the firing phase for adjustment. After outputting the start signal, the CPU 91 advances the process to step S180. The process of step S170 is a signal output process.

[0048] In step S180, the CPU 91 starts continuing the power factor adjustment. That is, thereafter, the CPU 91 repeats outputting a control signal to the thyristor based on the adjustment firing phase acquired in step S160. At this time, the CPU 91 grasps the phase of the current flowing through the power factor adjustment circuit 30 based on the switching state of the inverter 15, similar to step S170. Then, the CPU 91 outputs a control signal at an appropriate timing. After executing the process of step S180, the CPU 91 advances the process to step S190. The process of step S180 is a process for starting the adjustment continuation process.

[0049] In step S190, the CPU 91 judges whether the torque required for the motor generator 20 has decreased to less than the end value. The CPU 91 makes the judgment in step S190 based on the history of the torque required for the motor generator 20. Specifically, the CPU 91 acquires the latest required torque stored in the storage device 95, the required torque at the previous timing, and the end value. The end value is a value slightly lower than the above-mentioned specified torque N that defines the high torque state. The CPU 91 judges that the required torque has decreased to the end value if the following second judgment item is satisfied, and judges that the required torque has not decreased to the end value if not. The second judgment item is that the required torque at the previous timing is equal to or greater than the end value, and the latest required torque is less than the end value. If the second judgment item is not satisfied (step S190: NO), the CPU 91 performs the process of step S190 again. The CPU 91 repeats the process of step S190 until the second judgment item is satisfied. When the second check item is satisfied (step S190: YES), the CPU 91 proceeds to step S200. The situation in which the second check item is satisfied corresponds to a situation in which the accelerator pedal 60 is released. In other words, the accelerator operation amount ACC decreases to a value lower than the check value AK.

[0050] In step S200, the CPU 91 ends the power factor adjustment. That is, the CPU 91 ends the output of the control signal to the thyristor. After this, the CPU 91 ends the series of processes of the power factor adjustment process. After this, the CPU 91 executes the process of step S110 again.

[0051] <How to learn mapping> The map D is learned by the server 700 before being implemented in the vehicle 10. The server 700 collects data required for learning in advance when learning the map D. As shown in FIG. 3, the server 700 can communicate with a plurality of vehicles 10A having the same specifications as the vehicle 10 via an external communication line network 100. An event in which the accelerator operation amount ACC increases to a judgment value AK while the vehicle 10A is traveling is called a target event. The plurality of vehicles 10A sequentially transmit learning information U about the target event to the server 700. Specifically, when a target event occurs, the vehicle 10A acquires the values ​​of the following variables in the same manner as the acquisition process described above. The variables are the latest latitude Z1, the latest longitude Z2, the time identification value Z3, the month identification value Z4, and the weather identification value Z5. When the vehicle 10A acquires the values ​​of the variables, it transmits these variable values ​​and a success / failure discrimination value ZA indicating whether or not the motor generator 20 has entered a high torque state after a certain time T to the server 700 as learning information U. The success / failure discrimination value ZA is set to "1" if the motor generator 20 has entered a high torque state and "0" if the motor generator 20 has not entered a high torque state.

[0052] The server 700 generates a data set every time it acquires the learning information U. The data set is composed of teacher data and training data. The teacher data includes two variables, a true first probability and a true second probability. The training data includes five variables, a latest latitude Z1, a latest longitude Z2, a time identification value Z3, a month identification value Z4, and a weather identification value Z5. The server 700 treats the value of the success / failure identification value ZA in the learning information U as the true first probability. That is, if the motor generator 20 is in a high torque state, the true first probability is "1". If the motor generator 20 is not in a high torque state, the true first probability is "0". The server 700 sets the true second probability to a value obtained by replacing the values ​​of "0" and "1" in the true first probability. The server 700 treats the latest latitude Z1, the latest longitude Z2, the time identification value Z3, the month identification value Z4, and the weather identification value Z5 in the learning information U as training data.

[0053] When the server 700 accumulates the number of data sets necessary to learn the mapping D, it uses the multiple data sets to learn the mapping D. That is, the server 700 adjusts the input side coefficients wFjk and output side coefficients wSij of the mapping D so that the difference between the value output by the mapping D with the training data as input and the value of the teacher data becomes a predetermined value or less. Then, the server 700 determines that the learning is complete when the difference becomes a predetermined value or less.

[0054] <Operation of the embodiment> When the accelerator operation amount ACC increases to the judgment value AK (step S110: YES), the CPU 91 acquires the latest latitude Z1, the latest longitude Z2, the time identification value Z3, the month identification value Z4, and the weather identification value Z5 (step S120). Then, the CPU 91 inputs the values ​​of these variables and calculates a first probability R1, which is the probability that the motor generator 20 will be in a high torque state later (step S140). When the CPU 91 determines from the first probability R1 that the motor generator 20 will be in a high torque state later (step S150: YES), it acquires an adjustment firing phase and generates an initial control signal for the thyristor (step S160). At this time, the motor generator 20 has not yet entered a high torque state. That is, the CPU 91 completes the process required to start adjusting the power factor before the motor generator 20 enters a high torque state.

[0055] Thereafter, the CPU 91 outputs an initial control signal, i.e., a start signal, to the thyristor at a timing according to the firing phase for adjustment (step S170). When the thyristor is switched on in response to this control signal, adjustment of the power factor begins. Due to the relationship between the judgment value AK and the fixed time T, the torque of the motor generator 20 at this time is slightly lower than the specified torque N. In other words, the CPU 91 starts adjusting the power factor shortly before the motor generator 20 enters a high torque state. In other words, the CPU 91 starts adjusting the power factor without any delay with respect to the timing at which the motor generator 20 enters a high torque state.

[0056] Thereafter, the CPU 91 repeatedly outputs the control signal to the thyristor (step S180). In this way, the CPU 91 maintains the power factor between the inverter 15 and the motor generator 20 at a value close to 1 while the motor generator 20 is in a high torque state.

[0057] <Effects of the embodiment> (1) The CPU 91 uses a mapping D to predict the future state of the motor generator 20. The input variables of this mapping D include various variables that indicate the driving information of the vehicle 10. That is, the CPU 91 calculates the probability that the motor generator 20 will enter a high torque state based on not only the information on the accelerator operation amount ACC but also the driving information of the vehicle 10 when the accelerator operation amount ACC reaches the judgment value AK. In this way, by calculating the probability of entering a high torque state based on the driving information of the vehicle 10, the CPU 91 can accurately predict the future state of the motor generator 20. Then, the CPU 91 can accurately grasp that the motor generator 20 will enter a high torque state a certain time T after the accelerator operation amount ACC increases to the judgment value AK.

[0058] If it is possible to know that motor generator 20 will subsequently enter a high-torque state at a stage before the motor generator 20 does so, the following becomes possible. That is, CPU 91 can complete the preparation process for outputting a start signal before motor generator 20 enters the high-torque state. This allows CPU 91 to output the start signal before motor generator 20 enters the high-torque state. As a result, CPU 91 can start adjusting the power factor without any delay with respect to the timing at which motor generator 20 enters the high-torque state, and control the power factor to a value close to 1.

[0059] Here, when the torque of the motor generator 20 is high, the power consumption of the battery 11 is high. As described above, being able to control the power factor to a value close to 1 from the timing when the motor generator 20 enters a high torque state makes it possible to minimize reactive power in a situation where power consumption is high, from the initial timing when power consumption becomes high. This makes it possible to minimize the power consumption of the battery 11. As a result, it becomes possible to employ a battery 11 with a correspondingly small capacity.

[0060] (2) The input variables include a location variable that indicates the location where the vehicle 10 is traveling. Here, the roadway includes locations where the motor generator 20 is likely to enter a high torque state, such as uphill roads and merging roads onto a main highway. By including the location variable in the input variables, it is possible to calculate the probability of entering a high torque state by taking into account information on whether or not the location is likely to enter a high torque state. This increases the accuracy of calculating the probability.

[0061] (3) The input variables include a time variable that indicates the time when the vehicle 10 is traveling. Even when traveling at the same location, for example, when traveling at night, the vehicle speed SP will be lower and there will be fewer opportunities for sudden acceleration compared to when traveling during the day. By including the time variable in the input variables, it is possible to calculate the probability of a high torque state taking into account information on whether or not the time is a time period when a high torque state is likely to occur. Therefore, the accuracy of calculating the probability is improved.

[0062] (4) The input variables include a seasonal variable indicating the season in which the vehicle 10 is traveling. For example, the road surface may be frozen in winter. Therefore, even if traveling at the same location at the same time, the vehicle speed SP may be lower in winter than in other seasons, and the opportunities for sudden acceleration may be fewer. In other words, in winter, there is a tendency similar to that of the nighttime. By including the seasonal variable in the input variables, the probability of a high torque state occurring can be calculated taking into account information on whether or not it is a season in which a high torque state is likely to occur. Therefore, the calculation accuracy of the probability is increased.

[0063] (5) The input variables include weather variables that indicate the weather at the location where the vehicle 10 is traveling. Even if traveling at the same location at the same time, for example, when it is raining, the vehicle speed SP will be lower than when it is sunny, and the opportunities for sudden acceleration may be fewer. When it is snowing, the opportunities for sudden acceleration may be even fewer. By including weather variables in the above input variables, the probability of a high torque state can be calculated taking into account information on whether the weather is likely to result in a high torque state. Therefore, the accuracy of calculating the probability is improved.

[0064] <Example of change> This embodiment can be modified as follows: This embodiment and the following modifications can be combined with each other to the extent that no technical contradiction occurs.

[0065] The mapping D learned by the server 700 may be implemented in the vehicle 10A, and then the mapping D may be updated. That is, even after the mapping D is implemented in the vehicle 10A, each vehicle 10A transmits learning information U to the server 700 as needed. The server 700 uses the learning information U as teacher data and training data to update the mapping D, for example, on a monthly basis. The server 700 then transmits the updated mapping D to the vehicle 10A. In this way, the mapping D that reflects the latest trends can always be used.

[0066] In the above embodiment, a unified mapping D was created for multiple vehicles 10A of the same specifications. However, a mapping D may be created for each vehicle 10A. If learning information U is accumulated for each vehicle 10A and learning is performed, a mapping D can be created for each vehicle 10A. The points and times at which the motor generator 20 is likely to enter a high torque state may differ for each vehicle 10A, i.e., for each occupant who uses the vehicle 10A. In this modified example, a mapping D can be created that takes into account the characteristics of each occupant.

[0067] When creating the mapping D for each vehicle 10A, the following adjustments may be made. For example, the number of high torque occurrences, which is the number of times the motor generator 20 entered a high torque state during a certain period of time, such as one month, is measured. When the number of high torque occurrences is large, the mapping D may be adjusted so that the value of the first probability R1 is higher than when the number of high torque occurrences is small. For example, when the number of high torque occurrences is large, at the time when the probabilistic prototypes y(1) to y(2), which are the output of the neural network, are calculated, the probabilistic prototype y(1) may be multiplied by a coefficient that increases the value of the probabilistic prototype y(1).

[0068] When creating a mapping D for each vehicle 10A as in the above modification, the mapping D may be learned by a control device of the vehicle 10A instead of the server 700. In this case, the learning information U may be stored in a storage device of the vehicle 10A.

[0069] The method of determining the fixed time T and the judgment value AK is not limited to the example of the above embodiment. The judgment value AK and the fixed time T may be set so that the relationship that the motor generator 20 enters the high torque state after the fixed time T from when the accelerator operation amount ACC becomes the judgment value AK is satisfied. For example, the fixed time T and the judgment value AK may be determined for each vehicle 10A. In this case, it is possible to set the judgment value AK based on the operation characteristics of each occupant, for example, by lowering the judgment value AK for the vehicle 10A of an occupant who depresses the accelerator pedal 60 at a high speed and by raising the judgment value AK for the vehicle 10A of an occupant who depresses the accelerator pedal 60 at a slow speed. If the judgment value AK and the fixed time T are changed from those in the above embodiment, the mapping D may be learned using the learning information U obtained with the changed judgment value AK and the fixed time T as the setting conditions. In this way, the mapping D targeting the changed judgment value AK and the fixed time T can be obtained. Note that when changing the fixed time T, the required time must be taken into consideration. As described in the above embodiment, if the fixed time is longer than the required time, the start signal can be reliably transmitted before the motor generator 20 enters the high torque state.

[0070] The method of determining the specified torque N that specifies the high torque state is not limited to the example in the above embodiment. The specified torque N may be set appropriately in accordance with the state of the motor generator 20 that is to be the subject of grasping the future state using the mapping D. If the specified torque N is changed from that in the above embodiment, the mapping D may be learned using the learning information U that is obtained with the changed specified torque N as a setting condition. In this way, the mapping D targeted for the changed specified torque N may be obtained.

[0071] The variables used as the point variables are not limited to those in the above embodiment. For example, the latitude may be divided into certain ranges, and a value indicating a representative point of each range may be used as the point variable indicating the latitude. Similarly, the longitude may be divided into certain ranges, and a value indicating a representative point of each range may be used as the point variable indicating the longitude. The point variable may be any variable that indicates the point at which the vehicle 10 is traveling.

[0072] The variables used as the time variables are not limited to the examples in the above embodiment. For example, a day may be divided into six-hour periods, and an identification value may be set for each divided time period. That is, time period identification values ​​indicating each time period of "morning," "afternoon," "evening," and "late night" may be used as the time variables. The time variable may be anything that indicates the time when the vehicle 10 is traveling.

[0073] The variables used as seasonal variables are not limited to the examples in the above embodiment. For example, a year may be divided into three-month periods, and an identification value may be set for each set of months. In other words, the season identification values ​​indicating the seasons of "spring," "summer," "autumn," and "winter" may be used as seasonal variables.

[0074] The variables used as the weather variables are not limited to those in the above embodiment. For example, the probability of precipitation at the location where the vehicle 10 is traveling may be used as the weather variable. The weather variable may be anything that indicates the weather at the location where the vehicle 10 is traveling.

[0075] The variables used as variables indicating the travel information of the vehicle 10 are not limited to the examples of the above embodiment. Here, as described above, examples of the travel information of the vehicle 10 include geographical information of the point where the vehicle 10 is traveling, time information, and information on the surrounding environment such as weather. Information indicating the surrounding environment includes not only the weather but also temperature reflecting the icy state of the road surface. In addition, examples of the travel information of the vehicle 10 include information on the travel road at the point where the vehicle 10 is traveling. That is, a travel road variable indicating information on the travel road on which the vehicle 10 is traveling may be used as a variable indicating the travel information of the vehicle 10. Specifically, an identification value indicating the type of road, such as an expressway, a general national road, or a prefectural road, may be used as the travel road variable. In addition, the gradient of the travel road at the point where the vehicle 10 is traveling may be used as the travel road variable.

[0076] The input variables are not limited to the examples of the above embodiment. Other input variables may be adopted instead of or in addition to those shown in the above embodiment. The number of input variables may be reduced from that of the above embodiment. The number of input variables may be two or more. The input variables may include at least one variable indicating the driving information of the vehicle 10. In other words, variables other than the variable indicating the driving information of the vehicle 10 may be adopted as the input variables. If the variable is useful for determining whether the motor generator 20 will be in a high torque state a certain time after the operation variable reaches the judgment value, the accuracy of the value of the output variable can be improved.

[0077] The output variable is not limited to the example of the above embodiment. The output variable may be any variable that indicates whether the motor generator 20 will enter a high torque state a certain time after the operation variable reaches a judgment value. For example, the output variable may be a current value that flows through the motor generator 20 a certain time after the operation variable reaches a judgment value. In this case, if this current value is higher than a current value corresponding to a specified torque N, it can be determined that the motor generator 20 will enter a high torque state.

[0078] The number of output variables is not limited to the example in the above embodiment. For example, three output variables, the first probability R1, the second probability R2, and the current value, may be output. The configuration of the mapping D is not limited to the example of the above embodiment. For example, the number of intermediate layers in the neural network may be two or more.

[0079] The overall configuration of the vehicle 10 is not limited to the example of the above embodiment. For example, the vehicle 10 may be equipped with an internal combustion engine as well as the motor generator 20 as a drive source. The vehicle 10 may also have a continuously variable transmission. The continuously variable transmission and the motor generator 20 may be connected to each other.

[0080] The variable adopted as the manipulated variable is not limited to the accelerator operation amount ACC. The manipulated variable may be one whose value changes in response to the operation of the occupant. For example, as in the above modification, if the vehicle 10 is equipped with a continuously variable transmission, the gear ratio of the continuously variable transmission may be adopted as the manipulated variable. The gear ratio of the continuously variable transmission changes in value in response to the depression of the accelerator pedal 60, i.e., an acceleration request. When the above gear ratio is adopted as the manipulated variable, the judgment value, etc. may be changed to a value suitable for the above gear ratio. That is, the judgment value may be set in consideration of the fixed time so that the relationship that the motor generator 20 becomes in a high torque state after a fixed time after the above gear ratio reaches the judgment value is satisfied. In addition, the manipulated variable may be the vehicle speed SP of the vehicle 10. The vehicle speed SP also changes in value in response to the operation of the accelerator pedal 60 by the occupant. Therefore, the judgment value may be set so that the relationship that the motor generator 20 becomes in a high torque state after a fixed time after the vehicle speed SP reaches the judgment value is satisfied.

[0081] When the mapping D is created for each vehicle 10A, it is also possible to set an operation variable for each vehicle 10A. That is, if there is an operation that the occupant is likely to perform on the vehicle 10A before putting the motor generator 20 into a high torque state, a variable whose value changes according to the operation may be used as the operation variable. For example, it is assumed that a certain occupant has a tendency to put the motor generator 20 into a high torque state when driving on a straight road after a curve. In this case, the steering angle of the steering wheel may be used as the operation variable. The judgment value may be a value at which the steering wheel is in a neutral position. The output variable of the mapping D may be the probability that the motor generator 20 will be in a high torque state a certain time after the steering angle of the steering wheel returns to the neutral position.

[0082] The configuration of the power factor adjustment circuit 30 is not limited to the example of the above embodiment. The power factor adjustment circuit 30 may be configured to adjust the power factor between the inverter 15 and the motor generator 20. In other words, a variable capacitor may be connected in parallel to the coil of the motor generator 20. For example, a variable capacitor capable of adjusting the area of ​​two opposing electrode pairs may be used. Even if the configuration of the power factor adjustment circuit 30 is changed, when it is determined that the motor generator 20 will be in a high torque state, a pre-processing required to start power factor adjustment may be performed, and a start signal to start power factor adjustment may be output to the power factor adjustment circuit 30. [Explanation of symbols]

[0083] 10. Vehicle 11. Battery 15…Inverter 20...Motor generator 30...Power factor adjustment circuit 91…CPU 93...ROM 95…Storage device

Claims

[Claim 1] A battery; A motor that receives power from the battery; an inverter that converts DC to AC between the battery and the motor; a power factor adjustment circuit for adjusting a power factor of power between the inverter and the motor; The present invention is applied to an electric vehicle having A storage device and an execution device, the storage device pre-stores mapping data that defines a mapping that has been learned by machine learning and that outputs a value of an output variable when values ​​of a plurality of input variables are input; the plurality of input variables are variables indicating travel information of the electric vehicle when a value of an operation variable, the value of which changes in response to an operation of the electric vehicle by a passenger, reaches a predetermined determination value; the output variable is a variable indicating whether or not the motor will be in a high torque state a certain time after the operation variable reaches the determination value, The execution device is a first determination process for determining whether or not the value of the operation variable has reached the determination value; an acquisition process of acquiring values ​​of the plurality of input variables when it is determined in the first determination process that the value of the operation variable has reached the determination value; a calculation process for calculating a value of the output variable by inputting the values ​​of the input variables acquired by the acquisition process to the mapping; a second determination process for determining whether or not the motor is in a high torque state based on the value of the output variable calculated in the calculation process; a signal output process of outputting, to the power factor adjustment circuit, a start signal for starting adjustment of the power factor so that the power factor approaches 1 when it is determined in the second determination process that the motor will be in a high torque state; Run A control device for an electric vehicle.

Images