Failure prediction system, failure prediction method, and failure prediction program

Abstract

In the present invention, an acquisition unit acquires traveling data of an electric mobile body. A prediction unit predicts an aging failure of an electric machine conversion unit including a motor for driving a drive wheel of the electric mobile body and a drive circuit thereof, on the basis of the traveling data of the electric mobile body. The traveling data includes an input voltage of the drive circuit, input current of the drive circuit, the rotation speed of the motor driven by the drive circuit, and the rotation torque of the motor. The prediction unit predicts an aging failure of the electric machine conversion unit on the basis of transition of a value statistically indicating a relationship between input power of the drive circuit based on the input voltage and the input current of the drive circuit and a shaft output of the motor based on the rotation speed and the rotation torque of the motor.

Description

Technical Field

[0001] The present disclosure relates to a failure prediction system, a failure prediction method, and a failure prediction program for predicting the aging failures of an inverter and a motor mounted on an electric moving body.

Background Art

[0002] Electric vehicles (EVs) have been spreading mainly for commercial vehicles such as delivery vehicles. In recent years, the running data of EVs (such as battery information and vehicle control information) has been stored on the cloud, and an environment where it can be utilized in various fields is being constructed.

[0003] In an EV, a motor, an inverter, and a battery pack are key devices. The power elements used in the inverter (for example, MOSFET (METAL-OXIDE SEMICONDUCTOR FIELD-EFFECT TRANSMITTER), IGBT (INSULATED GATE BIPOLAR TRANSISTOR)) deteriorate over time. A major factor in the deterioration of the power elements is an increase in the contact resistance of the bonding wires. This is caused by metal fatigue due to heat cycles, and an increase in the contact resistance of the bonding wires appears as an increase in the loss of the power elements (efficiency decrease). Patent Document 1 discloses a method for predicting the life of the IGBT of an inverter from the difference between the input power and the output power of the inverter.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

[0005] Patent Document 1 relates to an inverter for driving a crane motor. However, in the case of an inverter that drives a motor that rotates at high speed, such as in an EV, the voltage and current of the three-phase sinusoidal AC between the inverter and the motor change rapidly. In this case, the method for calculating the difference between the input power and output power of the inverter requires high-speed sampling of the inverter's input voltage, input current, output voltage, and output current in order to ensure the correspondence between the inverter's input power and output power. Storing this log data requires high-speed access and large-capacity memory. Using such high-spec memory results in high costs.

[0006] By the way, predicting the aging degradation of components other than power elements, such as electrolytic capacitors, coils, and fans, requires dedicated sensors for each. Therefore, predicting the aging degradation of these components installed in EVs requires a design change to add dedicated sensors. On the other hand, the aging degradation of power elements can be predicted without adding dedicated sensors if the losses can be predicted.

[0007] This disclosure is made in light of these circumstances, and its purpose is to provide a technology for predicting the aging deterioration of the electromechanical conversion section of an electric mobile vehicle at low cost.

[0008] To solve the above problems, a failure prediction system in one aspect of the present disclosure includes an acquisition unit that acquires driving data of an electric mobile body, and a prediction unit that predicts age-related failures of an electromechanical conversion unit including a motor that drives the drive wheels of the electric mobile body and its drive circuit, based on the driving data of the electric mobile body. The driving data includes the input voltage of the drive circuit, the input current of the drive circuit, the rotational speed of the motor driven by the drive circuit, and the rotational torque of the motor. The prediction unit predicts age-related failures of the electromechanical conversion unit based on the changes in values ​​that statistically show the relationship between the input power of the drive circuit based on the input voltage and input current of the drive circuit and the shaft output of the motor based on the rotational speed and rotational torque of the motor.

[0009] Furthermore, any combination of the above components, as well as conversions of the expressions of this disclosure between devices, systems, methods, computer programs, recording media on which computer programs are recorded, etc., are also valid as aspects of this disclosure.

[0010] According to this disclosure, it is possible to predict the deterioration of the electromechanical conversion section of an electric mobile device over time at low cost. [Brief explanation of the drawing]

[0011] [Figure 1] This figure shows a schematic configuration of an electric vehicle according to an embodiment. [Figure 2] This diagram shows a schematic configuration of the drive system for an electric vehicle. [Figure 3] This figure shows an example configuration of a failure prediction system according to an embodiment. [Figure 4] This figure shows an example of a graph obtained by plotting multiple data points showing the losses in the electromechanical converter of a certain electric vehicle over a specified period and performing a linear regression. [Figure 5] This figure shows an example of a graph obtained by plotting multiple data points showing the losses in the electromechanical converter of the same electric vehicle over a reference period and performing a linear regression. [Figure 6] This figure schematically shows the change in the slope of the regression line. [Figure 7A] Figure 7A is a plot of multiple data points that show the relationship between the speed of the electric vehicle and the input power of the inverter, generated from the driving data that forms the basis of the graph shown in Figure 5. [Figure 7B] Figure 7B is a plot of multiple data points showing the correspondence between the speed of the electric vehicle and the shaft output of the motor, generated from the driving data that forms the basis of the graph shown in Figure 5. [Figure 8] This flowchart shows the processing flow for predicting age-related failures of the electromechanical converter unit using the failure prediction system according to the embodiment. [Modes for carrying out the invention]

[0012] Figure 1 is a diagram showing the schematic configuration of an electric vehicle 3 according to an embodiment. In this embodiment, the electric vehicle 3 is assumed to be a pure EV without an internal combustion engine. The electric vehicle 3 shown in Figure 1 is a rear-wheel drive (2WD) EV equipped with a pair of front wheels 31F, a pair of rear wheels 31R, and a motor 34 as a power source. The pair of front wheels 31F are connected by a front axle 32F, and the pair of rear wheels 31R are connected by a rear axle 32R. The transmission 33 transmits the rotation of the motor 34 to the rear axle 32R at a predetermined conversion ratio. Note that the electric vehicle 3 may also be a front-wheel drive (2WD) or 4WD vehicle.

[0013] The power supply system 40 comprises a battery pack 41 and a management unit 42, the battery pack 41 containing multiple cells. Lithium-ion battery cells, nickel-metal hydride battery cells, etc., can be used as the cells. Hereinafter, this specification assumes the use of lithium-ion battery cells (nominal voltage: 3.6-3.7V). The management unit 42 monitors the voltage, temperature, current, SOC (STATE OF CHARGE), and SOH (STATE OF HEALTH) of the multiple cells contained in the battery pack 41 and transmits this information to the vehicle control unit 30 via the in-vehicle network. For example, CAN (CONTROLLER AREA NETWORK) or LIN (LOCAL INTERCONNECT NETWORK) can be used as the in-vehicle network.

[0014] The inverter 35 is a drive circuit that drives the motor 34. During acceleration, it converts the DC power supplied from the battery pack 41 into AC power and supplies it to the motor 34. During regeneration, it converts the AC power supplied from the motor 34 into DC power and supplies it to the battery pack 41. During acceleration, the motor 34 rotates in accordance with the AC power supplied from the inverter 35. During regeneration, it converts the rotational energy due to deceleration into AC power and supplies it to the inverter 35.

[0015] FIG. 2 is a diagram showing a schematic configuration of a drive system of an electric vehicle 3. In FIG. 2, an example is shown in which a three-phase AC motor is used as a motor 34 for driving the electric vehicle 3, and the three-phase AC motor 34 is driven by a three-phase inverter 35. The three-phase inverter 35 converts DC power supplied from the battery pack 41 into three-phase AC power with phases shifted by 120 degrees each, and drives the three-phase AC motor 34.

[0016] The inverter 35 includes a first arm in which a first switching element Q1 and a second switching element Q2 are connected in series, a second arm in which a third switching element Q3 and a fourth switching element Q4 are connected in series, and a third arm in which a fifth switching element Q5 and a sixth switching element Q6 are connected in series. The first to third arms are connected in parallel to the battery pack 41.

[0017] In FIG. 2, IGBTs are used for the first switching element Q1 to the sixth switching element Q6. The first diodes D1 to the sixth diodes D6 are connected in anti-parallel to the first switching element Q1 to the sixth switching element Q6, respectively. When MOSFETs are used for the first switching element Q1 to the sixth switching element Q6, parasitic diodes formed in the source-to-drain direction are used as the first diodes D1 to the sixth diodes D6.

[0018] The motor controller 36 acquires the input DC voltage and input DC current of the inverter 3 35 detected by the input voltage / current sensor 381, the output AC voltage and output AC current of the inverter 35 detected by the output voltage / current sensor 382, and the rotational speed and rotational torque of the three-phase AC motor 34 detected by the rotational speed / torque sensor 383. Further, the motor controller 36 acquires an acceleration signal or a brake signal according to the operation of the driver or generated by the automatic driving controller.

[0019] Based on these input parameters, the motor controller 36 generates a PWM signal for driving the inverter 35 and outputs it to the gate driver 37. The gate driver 37 generates drive signals for the first switching element Q1 - the sixth switching element Q6 based on the PWM signal input from the motor controller 36 and a predetermined carrier wave, and inputs them to the gate terminals of the first switching element Q1 - the sixth switching element Q6.

[0020] The motor controller 36 transmits the input DC voltage of the inverter 35, the input DC current of the inverter 35, the rotational speed of the motor 34, and the rotational torque of the motor 34 to the vehicle control unit 30 via the in - vehicle network.

[0021] Returning to FIG. 1. The vehicle control unit 30 is a vehicle ECU (ELECTRONIC CONTROL UNIT) that controls the entire electric vehicle 3 and may be composed of, for example, an integrated VCM (VEHICLE CONTROL MODULE).

[0022] The vehicle speed sensor 385 generates a pulse signal proportional to the rotational speed of the front wheel axle 32F or the rear wheel axle 32R and transmits the generated pulse signal to the vehicle control unit 30. The vehicle control unit 30 detects the speed of the electric vehicle 3 based on the pulse signal received from the vehicle speed sensor 385.

[0023] The wireless communication unit 39 performs signal processing for wireless connection to the network via the antenna 39A. As a wireless communication network to which the electric vehicle 3 can be wirelessly connected, for example, a mobile phone network (cellular network), a wireless LAN, V2I (VEHICLE - TO - INFRASTRUCTURE), V2V (VEHICLE - TO - VEHICLE), an ETC system (ELECTRONIC TOLL COLLECTION SYSTEM), DSRC (DEDICATED SHORT RANGE COMMUNICATIONS) can be used.

[0024] The vehicle control unit 30 can transmit driving data in real time to a cloud server or the company's own server for data storage using the wireless communication unit 39 while the electric vehicle 3 is running. The driving data includes the vehicle speed of the electric vehicle 3, the voltage, current, temperature, state of charge (SOC), state of heat (SOH) of multiple cells contained in the battery pack 41, the input DC voltage of the inverter 35, the input DC current of the motor 34, and the rotational speed and rotational torque of the motor 34. The vehicle control unit 30 periodically samples this data (for example, at 10-second intervals) and transmits it to the cloud server or the company's own server each time.

[0025] The vehicle control unit 30 may store the driving data of the electric vehicle 3 in its internal memory and transmit the stored driving data in a batch at a predetermined timing. For example, the vehicle control unit 30 may transmit the stored driving data to the terminal device at the sales office in a batch after the end of business for the day. The terminal device at the sales office transmits the driving data of multiple electric vehicles 3 to a cloud server or the company's own server at a predetermined timing.

[0026] Furthermore, when charging from a charger equipped with network communication functionality, the vehicle control unit 30 may transmit the driving data stored in its memory to the charger in a batch via the charging cable. The charger then transmits the received driving data to a cloud server or its own server. This example is effective for electric vehicles 3 that do not have wireless communication functionality.

[0027] Figure 3 shows an example configuration of the failure prediction system 10 according to an embodiment. The failure prediction system 10 is constructed with one or more servers. For example, the failure prediction system 10 may be constructed with a single in-house server installed in a data center or in-house facility. Alternatively, the failure prediction system 10 may be constructed with a cloud server used based on a cloud service. Furthermore, the failure prediction system 10 may be constructed with multiple in-house servers distributed across multiple locations (data centers, in-house facilities). Additionally, the failure prediction system 10 may be constructed with a combination of a cloud server used based on a cloud service and an in-house server. Finally, the failure prediction system 10 may be constructed with multiple cloud servers based on contracts with multiple cloud service providers.

[0028] The fault prediction system 10 comprises a processing unit 11 and a storage unit 12. The processing unit 11 includes a driving data acquisition unit 111, a prediction unit 112, and a notification unit 113. The functions of the processing unit 11 can be realized through the cooperation of hardware resources and software resources, or solely through hardware resources. Hardware resources that can be used include a CPU, ROM, RAM, GPU (Graphics Processing Unit), ASIC (Application Specific Integrated Circuit), FPGA (Field Programmable Gate Array), and other LSIs. Software resources that can be used include operating systems and applications.

[0029] The storage unit 12 includes a driving data storage unit 121. The storage unit 12 includes a non-volatile recording medium such as an HDD (HARD DISK DRIVE) or SSD (SOLID STATE DRIVE) and stores various types of data.

[0030] The driving data acquisition unit 111 acquires driving data of the electric vehicle 3 via the network and stores the acquired driving data in the driving data storage unit 121. The prediction unit 112 reads the driving data of the target electric vehicle 3 for a certain period (for example, one month) stored in the driving data storage unit 121 and predicts age-related failures of the inverter 35 and motor 34 (hereinafter collectively referred to as the electromechanical conversion unit) of the electric vehicle 3. A detailed explanation follows below.

[0031] The prediction unit 112 calculates the input power EP [W] of the inverter 35 at each sample time based on the input DC voltage V [V] and input DC current I [A] of the inverter 35 at each sample time included in the read-out driving data (see Equation 1).

[0032] EP=V×I (Formula 1) The prediction unit 112 calculates the shaft output MP [W] of the motor 34 at each sample time based on the rotational speed N [RPM] and rotational torque T [N·M] of the motor 34 at each sample time included in the read-out driving data (see Equation 2).

[0033] MP=2Π×N×T / 60 (Formula 2) The prediction unit 112 performs regression analysis on multiple data points showing the correspondence between the input power EP of the inverter 35 and the shaft output MP of the motor 34, based on driving data over a certain period, and generates a regression line. For example, the least squares method can be used for linear regression. Since the difference between the input power EP of the inverter 35 and the shaft output MP of the motor 34 represents the sum of the losses of the inverter 35 and the motor 34, each data point represents the instantaneous value of the loss of the electromechanical conversion unit. Note that the regression analysis performed by the prediction unit 112 is not limited to simple linear regression, but may also be multiple linear regression.

[0034] Figure 4 shows an example of a graph obtained by plotting multiple data points showing the loss of the electromechanical converter in a certain electric vehicle 3 during a target period and performing a linear regression. Figure 5 shows an example of a graph obtained by plotting multiple data points showing the loss of the electromechanical converter in the same electric vehicle 3 during a reference period and performing a linear regression. The target period is a one-month period, and the reference period is the same month of the previous year.

[0035] In the graphs shown in Figures 4 and 5, the X-axis represents the shaft output MP of the motor 34, and the Y-axis represents the input power EP of the inverter 35. The first quadrant of the graph shows the plotted data when the motor 34 is in operation. During operation, the motor 34 rotates based on the power supplied from the battery pack 41 to the inverter 35. That is, there is a relationship between electrical input (input power EP is positive) and mechanical output (shaft output MP is positive).

[0036] The third quadrant of the graph shows the plotted data for motor 34 during regeneration. During regeneration, the rotational energy of motor 34 is recovered to the battery pack 41 via inverter 35. In other words, the relationship is mechanical input (shaft output MP is negative) → electrical output (input power EP is negative).

[0037] The prediction unit 112 calculates each plot data based on the input DC voltage, input DC current, motor speed, and motor torque of the inverter 35, which are sampled at the same time. The input power EP of the inverter 35 fluctuates according to the accelerator opening of the electric vehicle 3.

[0038] Strictly speaking, there is a time lag between the input power EP of the inverter 35 and its reflection in the shaft output MP of the motor 34. Therefore, plot data will occur where the change in the input power EP of the inverter 35 due to the change in accelerator opening is not reflected in the shaft output MP of the motor 34. This plot data will appear in the second quadrant (electrical input - mechanical input) and the fourth quadrant (electrical output - mechanical output). However, if the number of plot data is large, the impact of the plot data appearing in the second and fourth quadrants will be negligible.

[0039] The regression line generated from the multiple plotted data shown in Figure 4 is given by (Equation 3) below, and the regression line generated from the multiple plotted data shown in Figure 5 is given by (Equation 4) below.

[0040] Y=1.1023X+2126.8 (Formula 3) R 2 =0.972 Y=1.1123X+2272.8 (Formula 4) R 2 =0.9745 For any of the regression lines, the coefficient of determination R 2 The correlation coefficient R squared exceeds 0.97, indicating an extremely strong positive correlation between the input power EP of the inverter 35 and the shaft output MP of the motor 34. The prediction unit 112 predicts the timing of age-related failures in the electromechanical converter based on the change in the slope of the regression line. In the examples shown in Figures 4 and 5, the slope of the regression line increases from 1.1023 to 1.1123 in one year. This increase represents the increase in losses (efficiency decrease) in the electromechanical converter.

[0041] Figure 6 schematically shows the change in the slope of the regression line. The Y-intercept of the regression line represents the input power EP [W] of the inverter 35 when the shaft output MP of the motor 34 is 0 [W]. That is, it represents the fixed loss (offset) of the electromechanical converter. The value of the slope represents the conversion efficiency of the electromechanical converter. The prediction unit 112 determines that the time of age-related failure of the electromechanical converter is approaching if the slope of the regression line for the target period increases by a threshold or more compared to the slope of the regression line for the reference period. This threshold changes depending on the allowable losses of the motor 34 and the inverter 35, but for example, it may be set to a value that is an increase of a predetermined value (e.g., 1.0%) from the initial value of the slope of the regression line.

[0042] In the graphs shown in Figures 4-6, an example is shown where the X-axis represents the shaft output MP of the motor 34 and the Y-axis represents the input power EP of the inverter 35. However, it is also possible to use the input power EP of the inverter 35 as the X-axis and the Y-axis as the shaft output MP of the motor 34. In this case, the fixed loss of the electromechanical converter appears in the X-intercept of the regression line, and the conversion efficiency of the electromechanical converter appears in the slope of the regression line. The prediction unit 112 determines that the time of age-related failure of the electromechanical converter is approaching if the slope of the regression line for the target period falls by a threshold or more compared to the slope of the regression line for the reference period.

[0043] When it is determined that the electromechanical converter is nearing the time of age-related failure, the notification unit 113 sends an alert to the electric vehicle 3 on which the electromechanical converter is installed, or to the operation management terminal device (not shown) managing the electric vehicle 3, indicating that the failure of the electromechanical converter is imminent.

[0044] Upon receiving the alert notification, the user takes the affected electric vehicle 3 to a dealer or repair shop to undergo a detailed fault diagnosis of the inverter 35 and motor 34. Based on this detailed fault diagnosis, the user can schedule repairs or replacements of the inverter 35 or motor 34, or make an appointment for replacement after a specified period.

[0045] Furthermore, if the failure prediction system 10 can acquire detection values ​​from vibration sensors installed on the bearings of the motor 34, the prediction unit 112 can estimate the wear of the motor 34 bearings based on the detection values ​​from the vibration sensors and estimate the increase in motor 34 losses. The prediction unit 112 can then estimate the increase in inverter 35 losses by subtracting the increase in motor 34 losses from the increase in losses of the electromechanical converter. In this case, the prediction unit 112 can determine when the inverter 35 will experience age-related failures.

[0046] Incidentally, the conversion efficiency of motor 34 varies depending on the operating point, which is defined by the combination of rotational speed N [RPM] and rotational torque T [N·M]. Efficiency decreases whether the rotational speed N [RPM] increases or decreases from the optimal rotational speed. Basically, the further the rotational speed deviates from the optimal speed, the lower the efficiency becomes. Similarly, efficiency decreases whether the rotational torque T [N·M] increases or decreases from the optimal torque. Basically, the further the torque deviates from the optimal torque, the lower the efficiency becomes. For example, in a certain motor, the operating point with maximum efficiency (94%) is around 2300 [RPM] for rotational speed N and around 90 [N·M] for rotational torque T. Note that the efficiency map of a motor varies from model to model.

[0047] Figure 7A is a plot of multiple data points showing the correspondence between the speed of the electric vehicle 3 and the input power EP of the inverter 35, generated from the driving data that forms the basis of the graph shown in Figure 5. Figure 7B is a plot of multiple data points showing the correspondence between the speed of the electric vehicle 3 and the shaft output MP of the motor 34, generated from the driving data that forms the basis of the graph shown in Figure 5.

[0048] The data shown in Figures 7A and 7B only includes data where the speed of the electric vehicle 3 is within 88 [KM / H], indicating that the speed of the electric vehicle 3 is limited to within 88 [KM / H]. In the graph shown in Figure 7B, many data points are plotted along two lines R1 and R2 in the negative region (regenerative region) of the motor 34's shaft output MP. This indicates that the motor controller 36 of the electric vehicle 3 controls at least one of the motor 34's rotational speed N and rotational torque T during regeneration so that power is generated at the operating point where the motor 34's conversion efficiency is high. In the example shown in Figure 7B, control is performed in two regenerative modes; the first line R1 shows data when controlled in weak regenerative mode, and the second line R2 shows data when controlled in strong regenerative mode.

[0049] The conversion efficiency of motor 34 is approximately the same when controlled in weak regenerative mode. Similarly, the conversion efficiency of motor 34 is approximately the same when controlled in strong regenerative mode. In the regenerative region, a lot of data is plotted along the two lines R1 and R2, which means that the data in the regenerative region shows little variation in the conversion efficiency of motor 34.

[0050] The smaller the variation in the conversion efficiency of the motor 34, the more accurately the losses of the inverter 35 can be estimated from the losses of the electromechanical conversion unit. From this perspective, the prediction unit 112 may generate the regression line using only the driving data from the target period in a regenerative state where regenerative current flows from the motor 34 to the inverter 35. Similarly, the regression line for the reference period may be generated using only the driving data in a regenerative state. Furthermore, the prediction unit 112 may generate the regression line using only the driving data along the first line R1 or the driving data along the second line R2.

[0051] As a general principle in statistical processing, the larger the sample size, the more accurate the regression. From this perspective, it is desirable for the prediction unit 112 to generate the regression line using both the driving data from the powered state, when power current flows from the inverter 35 to the motor 34, and the driving data from the regenerative state, from the driving data for the target period. In particular, if sufficient driving data from the regenerative state cannot be secured, it is desirable to use driving data from both the powered state and the regenerative state. Therefore, if the driving data that should be used as the basis for generating the regression line is insufficient in either the powered state or the regenerative state, that is, if it is necessary to increase the amount of driving data that should be used as the basis for generating the regression line, it is preferable for the prediction unit 112 to switch to generating the regression line based on driving data from both the powered state and the regenerative state.

[0052] Furthermore, when sampling driving data to be used as the basis data for generating the regression line, it is preferable for the prediction unit 112 to weight the driving data. For example, a difference in weighting may be applied to driving data in the regenerative state, prioritizing driving data in the powered state over driving data in the regenerative state, or plotted data appearing in the second and fourth quadrants of the graphs shown in Figures 4 and 5 may be excluded from the driving data to be used as the basis data.

[0053] Figure 8 is a flowchart showing the process flow for predicting age-related failures of the electromechanical converter unit by the failure prediction system 10 according to the embodiment. The prediction unit 112 reads out the input DC voltage of the inverter 35, the input DC current of the inverter 35, the rotational speed of the motor 34, and the rotational torque stored in the driving data holding unit 121 for the target period (S10). The prediction unit 112 calculates the input power of the inverter 35 and the shaft output of the motor 34 for each sample time (S11). The prediction unit 112 performs a linear regression on multiple values ​​that show the correspondence between the input power of the inverter 35 and the shaft output of the motor 34 during the target period and calculates the slope of the regression line (S12).

[0054] The prediction unit 112 similarly calculates the slope of the regression line for the reference period. If the slope of the regression line for the reference period has already been calculated and that value is stored in the driving data storage unit 121, the prediction unit 112 reads that value and uses it.

[0055] The prediction unit 112 predicts the timing of failure of the electromechanical converter based on the slope of the regression line for the target period and the slope of the regression line for the reference period (S13). The notification unit 113 sends an alert to the electric vehicle 3 or the operation management terminal device (not shown) that manages the electric vehicle 3, if necessary.

[0056] As described above, this embodiment makes it possible to predict the aging deterioration of the electromechanical conversion unit of the electric vehicle 3 at low cost. By acquiring and saving the driving data of the electric vehicle 3, there is no need to add new parts to the electric vehicle 3 (for example, sensors for detecting failures of switching elements Q1-Q6). Failures of the electromechanical conversion unit can be predicted with high accuracy and low cost by analyzing log data alone.

[0057] Since the voltage and current of the three-phase sinusoidal AC between the inverter 35 and the motor 34 are not used, there is no need to save high-speed sampled log data. The input voltage and input current of the inverter 35 are DC, and the change in accelerator opening is not a high-speed change on the order of microseconds or milliseconds. Similarly, the change in rotational speed and rotational torque of the motor 34 is not a high-speed change on the order of microseconds or milliseconds. Therefore, there is little need to save high-speed sampled log data for the input voltage and input current of the inverter 35 and the rotational speed and rotational torque of the motor 34; it is sufficient to save low-speed sampled log data on the order of seconds.

[0058] Thus, in this embodiment, high-spec memory is not required, and no additional sensors are needed, so the additional hardware cost is basically zero. Prediction is possible using only existing cloud-stored data. Furthermore, since the focus is on the relationship between the input power of the inverter 35 and the shaft output of the motor 34 at the same time, it is possible to eliminate dependence on external factors such as the travel route and travel environment.

[0059] In this embodiment, by predicting the failure time of the electromechanical converter based on the predicted increase in losses over time, the user can be notified in advance and prompted to replace or repair the inverter 35 or motor 34. This avoids the inconvenience of being unable to drive due to a sudden failure of the inverter 35 or motor 34. The user can replace the inverter 35 or motor 34 at the optimal timing as a predictive maintenance measure. This allows the user to minimize downtime while pursuing economic rationality.

[0060] Furthermore, if the deterioration of the motor 34 can be estimated from existing vibration sensors, the prediction unit 112 can predict the timing of inverter 35 failure. By using driving data that is as close as possible to the efficiency of the motor 34 (for example, driving data only during regeneration), the prediction unit 112 can estimate the losses of the inverter 35 with higher accuracy.

[0061] The present disclosure has been explained above based on examples. The examples are illustrative, and it will be readily apparent to those skilled in the art that various modifications are possible in combinations of their components and processing steps, and that such modifications are also within the scope of the present disclosure.

[0062] The fault prediction system 10 described above may be implemented in the battery control unit 32 within the electric vehicle 3. In this case, a large memory capacity is required, but data loss can be minimized.

[0063] In the above embodiment, the electric vehicle 3 is assumed to be a four-wheeled electric vehicle. However, it may also be an electric motorcycle (electric scooter), electric bicycle, or electric kick scooter. Furthermore, electric vehicles include not only full-size electric vehicles but also low-speed electric vehicles such as golf carts and land cars used in shopping malls and entertainment facilities. In addition, the battery pack 41 is not limited to electric vehicle 3. For example, it can also be used on electric mobile devices such as electric boats, railway vehicles, and multicopters (drones).

[0064] The embodiments may be specified by the following items.

[0065] [Item 1] An acquisition unit (111) that acquires driving data of the electric mobile unit (3), The system includes a prediction unit (112) that predicts age-related failures of the electromechanical conversion unit (34, 35), which includes a motor (34) that drives the drive wheels (31R) of the electric mobile unit (3) and its drive circuit (35), based on the driving data of the electric mobile unit (3), The aforementioned driving data includes the input voltage of the drive circuit (35), the input current of the drive circuit (35), the rotational speed of the motor (34) driven by the drive circuit (35), and the rotational torque of the motor (34). The prediction unit (112) predicts age-related failures of the electromechanical conversion unit (34, 35) based on the changes in values ​​that statistically show the relationship between the input power of the drive circuit (35) based on the input voltage and input current of the drive circuit (35) and the shaft output of the motor (34) based on the rotational speed and rotational torque of the motor (34). A failure prediction system (10) characterized by the following.

[0066] According to this, the deterioration over time of the electromechanical conversion section (34, 35) of the electric mobile unit (3) can be predicted at low cost.

[0067] [Item 2] The prediction unit (112) predicts age-related failures of the electromechanical conversion unit (34, 35) based on the change in the slope of a regression line obtained by linearly regressing multiple data showing the correspondence between the input power of the drive circuit (35) and the shaft output of the motor (34) based on driving data over a certain period of time. The failure prediction system (10) described in item 1, characterized by the above.

[0068] According to this, it is possible to predict with high accuracy the changes in losses of the electromechanical conversion section (34, 35) over time.

[0069] [Item 3] The prediction unit (112) extracts driving data from the driving data within the specified period in which regenerative current is flowing from the motor (34) to the drive circuit (35), and generates the regression line. The failure prediction system (10) described in item 2, characterized by the above.

[0070] According to this, the losses of the electromechanical conversion unit (34, 35) can be estimated based on data with small variations in the efficiency of the motor (34).

[0071] [Item 4] The prediction unit (112) generates the regression line based on driving data from the driving period, specifically the driving data for both the state in which power current is flowing from the drive circuit (35) to the motor (34) and the state in which regenerative current is flowing from the motor (34) to the drive circuit (35). The failure prediction system (10) described in item 2, characterized by the above.

[0072] This allows us to secure a sufficient number of sample data to estimate the losses in the electromechanical conversion units (34, 35).

[0073] [Item 5] The prediction unit (112) generates the regression line based on either the driving data from the driving circuit (35) to the motor (34) during the specified period, or the driving data from the motor (34) to the drive circuit (35). When the prediction unit (112) increases the amount of driving data that should be used as the basis data for generating the regression line, it generates the regression line based on driving data from both the state in which power current is flowing from the drive circuit (35) to the motor (34) and the state in which regenerative current is flowing from the motor (34) to the drive circuit (35). The failure prediction system (10) described in item 2, characterized by the above.

[0074] This approach can prevent a decrease in prediction accuracy due to insufficient data.

[0075] [Item 6] Driving data from multiple electric mobile units (3) is stored in the server (12), The prediction unit (112) predicts age-related failures of the electromechanical conversion units (34, 35) based on the driving data stored in the server (12). A fault prediction system (10) according to any one of items 1 to 5, characterized by the above.

[0076] According to this, a cloud service can be realized that predicts and provides information on age-related failures of the electromechanical conversion unit (34, 35) based on driving data accumulated on the server (12).

[0077] [Item 7] Steps to acquire driving data of the electric mobile device (3), The method includes the step of predicting age-related failures of the electromechanical conversion unit (34, 35), which includes the motor (34) that drives the drive wheels (31R) of the electric mobile unit (3) and its drive circuit (35), based on the driving data of the electric mobile unit (3). The aforementioned driving data includes the input voltage of the drive circuit (35), the input current of the drive circuit (35), the rotational speed of the motor (34) driven by the drive circuit (35), and the rotational torque of the motor (34). The aforementioned prediction step predicts the age-related failure of the electromechanical conversion unit (34, 35) based on the changes in values ​​that statistically show the relationship between the input power of the drive circuit (35) based on the input voltage and input current of the drive circuit (35) and the shaft output of the motor (34) based on the rotational speed and rotational torque of the motor (34). A failure prediction method characterized by the following.

[0078] According to this, the deterioration over time of the electromechanical conversion section (34, 35) of the electric mobile unit (3) can be predicted at low cost.

[0079] [Item 8] Process for acquiring driving data of the electric mobile unit (3), Based on the driving data of the electric mobile body (3), the computer is instructed to perform a process to predict age-related failures of the electromechanical conversion unit (34, 35), which includes the motor (34) that drives the drive wheels (31R) of the electric mobile body (3) and its drive circuit (35). The aforementioned driving data includes the input voltage of the drive circuit (35), the input current of the drive circuit (35), the rotational speed of the motor (34) driven by the drive circuit (35), and the rotational torque of the motor (34). The aforementioned prediction process predicts age-related failures of the electromechanical conversion unit (34, 35) based on the changes in values ​​that statistically show the relationship between the input power of the drive circuit (35) based on the input voltage and input current of the drive circuit (35) and the shaft output of the motor (34) based on the rotational speed and rotational torque of the motor (34). A failure prediction program characterized by the following features.

[0080] According to this, the deterioration over time of the electromechanical conversion section (34, 35) of the electric mobile unit (3) can be predicted at low cost. [Explanation of Symbols]

[0081] 3 Electric vehicle, 10 Fault prediction system, 11 Processing unit, 111 Driving data acquisition unit, 112 Prediction unit, 113 Notification unit, 12 Storage unit, 121 Driving data holding unit, 30 Vehicle control unit, 31F Front wheel, 31R Rear wheel, 32F Front axle, 32R Rear axle, 33 Transmission, 34 Motor, 35 Inverter, 36 Motor controller, 37 Gate driver, 381 Input voltage / current sensor, 382 Output voltage / current sensor, 383 Rotation speed / torque sensor, 385 Vehicle speed sensor, 39 Wireless communication unit, 39A Antenna, 40 Power supply system, 41 Battery pack, 42 ​​Management unit, Q1, Q6 Switching elements, D1, D6 Diodes.

Claims

1. An acquisition unit that acquires driving data of an electric mobile device, The system includes a prediction unit that predicts age-related failures of the electromechanical conversion unit, which includes a motor that drives the drive wheels of the electric mobile unit and a drive circuit for the motor, based on the driving data of the electric mobile unit. The aforementioned driving data includes the input voltage of the drive circuit, the input current of the drive circuit, the rotational speed of the motor driven by the drive circuit, and the rotational torque of the motor. The prediction unit predicts age-related failures of the electromechanical converter based on the changes in values ​​that statistically represent the relationship between the input power of the drive circuit, which is based on the input voltage and input current of the drive circuit, and the shaft output of the motor, which is based on the rotational speed and rotational torque of the motor. A failure prediction system characterized by the following features.

2. The prediction unit predicts age-related failures of the electromechanical converter based on the changes in the slope of a regression line obtained by linearly regressing multiple data showing the correspondence between the input power of the drive circuit and the shaft output of the motor, based on driving data over a certain period. The failure prediction system according to feature 1.

3. The prediction unit extracts driving data from the driving data within the specified period in which regenerative current is flowing from the motor to the drive circuit, and generates the regression line. The failure prediction system according to claim 2.

4. The prediction unit generates the regression line based on driving data from the driving period, specifically, driving data from both the state in which power current is flowing from the drive circuit to the motor and the state in which regenerative current is flowing from the motor to the drive circuit. The failure prediction system according to claim 2.

5. The prediction unit generates the regression line based on either the driving data within the specified period in which a power current is flowing from the drive circuit to the motor, or the driving data in which a regenerative current is flowing from the motor to the drive circuit. When the prediction unit increases the amount of driving data that should be used as the basis data for generating the regression line, it generates the regression line based on driving data from both the state in which power current is flowing from the drive circuit to the motor and the state in which regenerative current is flowing from the motor to the drive circuit. The failure prediction system according to claim 2.

6. The driving data of multiple electric mobile units is stored in the server. The prediction unit predicts age-related failures of the electromechanical converter based on the driving data stored in the server. A failure prediction system according to any one of claims 1 to 5.

7. Steps to acquire driving data for an electric mobile device, The method includes the step of predicting age-related failures of an electromechanical conversion unit, including a motor that drives the drive wheels of the electric mobile unit and a drive circuit for the motor, based on the driving data of the electric mobile unit. The aforementioned driving data includes the input voltage of the drive circuit, the input current of the drive circuit, the rotational speed of the motor driven by the drive circuit, and the rotational torque of the motor. The aforementioned prediction step predicts the age-related failure of the electromechanical converter based on the changes in values ​​that statistically represent the relationship between the input power of the drive circuit, which is based on the input voltage and input current of the drive circuit, and the shaft output of the motor, which is based on the rotational speed and rotational torque of the motor. A failure prediction method characterized by the following.

8. The process of acquiring driving data for an electric mobile device, Based on the driving data of the electric mobile body, the computer is instructed to perform a process to predict age-related failures of the motor that drives the drive wheels of the electric mobile body and the electromechanical conversion unit including the motor's drive circuit. The aforementioned driving data includes the input voltage of the drive circuit, the input current of the drive circuit, the rotational speed of the motor driven by the drive circuit, and the rotational torque of the motor. The aforementioned prediction process predicts age-related failures of the electromechanical converter based on the changes in values ​​that statistically represent the relationship between the input power of the drive circuit, which is based on the input voltage and input current of the drive circuit, and the shaft output of the motor, which is based on the rotational speed and rotational torque of the motor. A failure prediction program characterized by the following features.

Images