Method for calculating torque of mechanical compressor for vehicle

Abstract

The present invention relates to a method for calculating the torque of a mechanical compressor for a vehicle, the method comprising: a sensor data receiving step of receiving sensor data measured by a sensor installed in an air conditioning device including the mechanical compressor for a vehicle; a first calculation step of calculating a theoretical torque value of the mechanical compressor for a vehicle from the sensor data through a predetermined relational expression; a second calculation step of calculating a torque correction value for the theoretical torque value from the sensor data through a pre-trained artificial neural network model; and a third calculation step of calculating the output torque of the mechanical compressor for a vehicle by adding the theoretical torque value and the torque correction value, and thus able to accurately calculate the output torque of the mechanical compressor for a vehicle. Accordingly, since the amount of refrigerant discharged from the mechanical compressor for a vehicle can be appropriately adjusted, cooling performance and efficiency are improved, vehicle fuel efficiency is improved, and engine stall can be prevented.

Description

Method for calculating torque of a mechanical compressor for vehicles

[0001] The present invention relates to a method for calculating torque of a mechanical compressor for a vehicle, and more particularly, to a method for calculating torque of a mechanical compressor for a vehicle that enables accurate calculation of the output torque of the mechanical compressor for a vehicle.

[0002] Typically, a vehicle includes a prime mover that generates the power required for operation, and the prime mover is formed by an internal combustion engine that converts chemical energy into mechanical energy, a motor that converts electrical energy into mechanical energy, etc.

[0003] Here, a vehicle that uses an internal combustion engine as a motor is referred to as an internal combustion engine vehicle, a vehicle that uses an internal combustion engine and a motor as a motor is referred to as a hybrid vehicle, and a vehicle that uses a motor as a motor is referred to as an electric vehicle.

[0004] Meanwhile, regardless of the type of engine, the vehicle includes an air conditioning device for cooling and heating the interior, and the air conditioning device is typically formed by a vapor compression refrigeration cycle. That is, the air conditioning device includes a compressor that receives power from the engine and compresses a low-temperature, low-pressure gaseous refrigerant into a high-temperature, high-pressure gaseous refrigerant, a first heat exchanger that cools (condenses) the high-temperature, high-pressure gaseous refrigerant compressed by the compressor into a low-temperature, high-pressure liquid refrigerant, an expansion unit that expands the low-temperature, high-pressure liquid refrigerant cooled by the first heat exchanger into a low-temperature, low-pressure liquid refrigerant, and a second heat exchanger that heats (evaporates) the low-temperature, low-pressure liquid refrigerant expanded by the expansion unit into a low-temperature, low-pressure gaseous refrigerant.

[0005] Here, the compressor is classified into a reciprocating compressor that compresses the refrigerant through the reciprocating motion of the piston according to the compression method, and a rotary compressor that performs compression while rotating, and the reciprocating compressor is classified into a crank-type compressor that transmits power to a plurality of pistons using a crank, a swash plate compressor that transmits power through a rotating shaft on which a swash plate is installed, etc. according to the power transmission method, and the rotary compressor can be classified into a vane rotary compressor that uses a rotating rotary shaft and vanes, and a scroll compressor that uses an orbiting scroll and a fixed scroll.

[0006] Additionally, the compressor may be classified into a mechanical compressor that uses the vehicle's engine and an electric compressor that uses a motor depending on the driving method.

[0007] Here, the swash plate type compressor is usually formed mechanically, and the swash plate type compressor formed mechanically and mounted on a vehicle is called a vehicle mechanical compressor. Recently, the vehicle mechanical compressor has been formed in a so-called variable capacity type that controls the refrigerant discharge amount by controlling the stroke of the piston by adjusting the inclination angle of the swash plate to improve compression performance and efficiency.

[0008] That is, a mechanical compressor for a vehicle includes a housing, a rotating shaft rotatably mounted on the housing, a pulley that rotates by receiving power from a vehicle engine, a clutch that selectively engages the pulley and the rotating shaft, a swash plate that is accommodated in a crank chamber of the housing and rotates together with the rotating shaft, a piston that reciprocates by the swash plate and sucks in and compresses refrigerant and discharges it to a discharge chamber of the housing, an inlet passage that guides refrigerant in the discharge chamber to the crank chamber so that an angle of inclination of the swash plate is adjusted to adjust a stroke of the piston, and a valve that adjusts an opening amount of the inlet passage.

[0009] And, the air conditioning device including the vehicle mechanical compressor calculates the output torque of the vehicle mechanical compressor through the air conditioning control unit, and controls the valve based on the calculated output torque to control the stroke (more precisely, the stroke of the piston) of the swash plate type vehicle mechanical compressor.

[0010] Here, the conventional air conditioning device also uses sensor data measured by a sensor installed in the air conditioning device to calculate the output torque of the mechanical compressor for the vehicle.

[0011] In this way, the refrigerant discharge amount of the vehicle mechanical compressor is varied as the stroke of the vehicle mechanical compressor is adjusted.

[0012] However, these conventional methods for calculating torque of mechanical vehicle compressors have the problem of inaccurately calculating output torque. Consequently, the refrigerant discharge rate of the mechanical vehicle compressor cannot be properly controlled, resulting in reduced cooling performance and efficiency, reduced fuel efficiency, and engine stalling.

[0013] Accordingly, the present invention aims to provide a method for calculating torque of a mechanical compressor for a vehicle capable of accurately calculating output torque.

[0014] The present invention provides a method for calculating torque of a mechanical compressor for a vehicle, comprising: a sensor data receiving step for receiving sensor data measured by a sensor installed in an air conditioning device including a mechanical compressor for a vehicle, in order to achieve the above-described object; a first calculation step for calculating a theoretical torque value of the mechanical compressor for a vehicle from the sensor data using a predetermined relational expression; a second calculation step for calculating a torque correction value for the theoretical torque value from the sensor data using a pre-learned artificial neural network model; and a third calculation step for calculating an output torque of the mechanical compressor for a vehicle by adding the theoretical torque value and the torque correction value.

[0015] The method further includes a count data generation step of generating count data for the driving time of the mechanical compressor for the vehicle, wherein the artificial neural network model can calculate the torque compensation value from the sensor data and the count data.

[0016] The above count data can be reset each time the air conditioner is turned on.

[0017] The method further includes a data normalization step of normalizing the sensor data and the count data to a value greater than or equal to 0 and less than or equal to 1, and the artificial neural network model can calculate the torque correction value from the normalized sensor data and count data.

[0018] The artificial neural network model can calculate a plurality of primary parameter values ​​from the sensor data and the count data, calculate at least one secondary parameter value by combining the plurality of primary parameter values, and calculate the torque correction value using the at least one secondary parameter value.

[0019] There may be at least four input variables input to the above artificial neural network model.

[0020] The above relationship is , TQbase is the theoretical torque value, vs is the volume per unit mass of the refrigerant of the air conditioning device, Ps is the suction pressure of the vehicle mechanical compressor, Pd is the discharge pressure of the vehicle mechanical compressor, is the mass flow rate of the refrigerant of the air conditioning device, Nc is the rotational speed per minute of the vehicle mechanical compressor, and n may be an eigentropic index.

[0021] The discharge pressure of the above vehicle mechanical compressor may be sensor data measured by the above sensor.

[0022] At least one of the volume per unit mass of the refrigerant of the air conditioning device, the suction pressure of the mechanical compressor for the vehicle, and the mass flow rate of the refrigerant of the air conditioning device may be a value calculated using the sensor data.

[0023] The suction pressure of the above-mentioned vehicle mechanical compressor may be a value calculated using the above-mentioned sensor data.

[0024] A method for calculating torque of a mechanical compressor for a vehicle according to the present invention comprises: a sensor data receiving step of receiving sensor data measured by a sensor installed in an air conditioning device including a mechanical compressor for a vehicle; a first calculation step of calculating a theoretical torque value of the mechanical compressor for a vehicle from the sensor data using a predetermined relational expression; a second calculation step of calculating a torque correction value for the theoretical torque value from the sensor data using a pre-learned artificial neural network model; and a third calculation step of calculating the output torque of the mechanical compressor for a vehicle by adding the theoretical torque value and the torque correction value; thereby accurately calculating the output torque of the mechanical compressor for a vehicle. Accordingly, the refrigerant discharge amount of the mechanical compressor for a vehicle can be appropriately controlled, thereby improving cooling performance and efficiency, enhancing vehicle fuel efficiency, and preventing engine stalling.

[0025] FIG. 1 is a flowchart illustrating a method for calculating torque of a mechanical compressor for a vehicle according to one embodiment of the present invention;

[0026] Figure 2 is a schematic diagram illustrating an artificial neural network model for calculating a torque compensation value in the vehicle mechanical compressor torque calculation method of Figure 1.

[0027] FIG. 3 is a schematic diagram illustrating a vehicle including an air conditioning device controlled by the vehicle mechanical compressor torque calculation method of FIG. 1;

[0028] Fig. 4 is a cross-sectional view showing a vehicle mechanical compressor included in the air conditioning device of Fig. 3;

[0029] Figure 5 is a diagram showing the accuracy of the output torque calculated by the vehicle mechanical compressor torque calculation method of Figure 1 in the form of a time series torque change.

[0030] Fig. 6 is a diagram showing the accuracy of the output torque calculated using the vehicle mechanical compressor torque calculation method of Fig. 1 in the form of a scatter plot and residual distribution.

[0031] Hereinafter, a method for calculating torque of a mechanical compressor for a vehicle according to the present invention will be described in detail.

[0032] FIG. 1 is a flowchart illustrating a method for calculating torque of a mechanical compressor for a vehicle according to an embodiment of the present invention, FIG. 2 is a system diagram illustrating an artificial neural network model for calculating a torque correction value in the method for calculating torque of a mechanical compressor for a vehicle of FIG. 1, FIG. 3 is a system diagram illustrating a vehicle including an air conditioning device controlled by the method for calculating torque of a mechanical compressor for a vehicle of FIG. 1, FIG. 4 is a cross-sectional view illustrating a mechanical compressor for a vehicle included in the air conditioning device of FIG. 3, FIG. 5 is a diagram illustrating the accuracy of an output torque calculated by the method for calculating torque of a mechanical compressor for a vehicle of FIG. 1 in the form of a time series torque change, and FIG. 6 is a diagram illustrating the accuracy of an output torque calculated by the method for calculating torque of a mechanical compressor for a vehicle of FIG. 1 in the form of a scatter plot and a residual distribution.

[0033] Referring to FIG. 3, a vehicle according to one embodiment of the present invention may include a motor (100) that generates power and an air conditioning device (200) that receives some of the power generated by the motor (100) to provide heating and cooling of the interior.

[0034] The above-mentioned engine (100) can be formed as an internal combustion engine that converts chemical energy into mechanical energy, a motor that converts electrical energy into mechanical energy, etc.

[0035] The air conditioning device (200) may be formed as a vapor compression refrigeration cycle. That is, the air conditioning device (200) may include a vehicle mechanical compressor (210) that receives power from the motor (100) to compress a low-temperature, low-pressure gaseous refrigerant into a high-temperature, high-pressure gaseous refrigerant, a first heat exchanger (220) that cools (condenses) the high-temperature, high-pressure gaseous refrigerant compressed by the vehicle mechanical compressor (210) into a low-temperature, high-pressure liquid refrigerant, an expansion unit (230) that expands the low-temperature, high-pressure liquid refrigerant cooled by the first heat exchanger (220) into a low-temperature, low-pressure liquid refrigerant, and a second heat exchanger (240) that heats (evaporates) the low-temperature, low-pressure liquid refrigerant expanded by the expansion unit (230) into a low-temperature, low-pressure gaseous refrigerant.

[0036] Here, referring to FIG. 4, the vehicle mechanical compressor (210) can be formed as a so-called variable capacity swash plate type. That is, the vehicle mechanical compressor (210) comprises a housing (211), a rotary shaft (212) rotatably mounted on the housing (211), a pulley (213) that receives power from the motor (100) and rotates, a clutch (214) that selectively engages the pulley (213) and the rotary shaft (212), a swash plate (215) that is accommodated in the crank chamber (V4) of the housing (211) and rotates together with the rotary shaft (212), a piston (216) that is inserted into the bore (211a) of the housing (211) and reciprocates by the swash plate (215) to suck in and compress refrigerant and discharge it into the discharge chamber (V3) of the housing (211), and a stroke of the piston (216) is controlled by adjusting the inclination angle of the swash plate (215) so as to guide the refrigerant in the discharge chamber (V3) to the crank chamber (V4). It may include an inlet passage (not shown), an exhaust passage (218) that guides the refrigerant in the crankcase (V4) to the suction chamber (V1) of the housing (211), and a valve (e.g., a mechanical pressure regulating valve or an electronic pressure regulating valve) (219) that controls the opening amount of the inlet passage (not shown).

[0037] In addition, the air conditioning device (200) further includes a sensor (250) for measuring the discharge pressure of the vehicle mechanical compressor (210) and an air conditioning control unit (270) for receiving sensor data measured by the sensor (250), calculating the output torque (TQout) of the vehicle mechanical compressor (210), and controlling the valve (219) based on the output torque (TQout), and the air conditioning control unit (270) can be configured to calculate the output torque (TQout) of the vehicle mechanical compressor (210) according to the vehicle mechanical compressor torque calculation method illustrated in FIG. 1.

[0038] Specifically, the vehicle mechanical compressor torque calculation method may include a sensor data receiving step (S1) of receiving the sensor data, and a first calculation step (S2) of calculating the theoretical torque value (TQbase) of the vehicle mechanical compressor (210) by substituting the sensor data received in the sensor data receiving step (S1) into a predetermined relational expression.

[0039] In addition, the vehicle mechanical compressor torque calculation method may further include a count data generation step (S3) of generating count data that counts the time elapsed since the vehicle mechanical compressor (210) is started (vehicle mechanical compressor operation time), a data normalization step (S4) of normalizing the sensor data received in the sensor data reception step (S1) and the count data generated in the count data generation step (S3) to a number greater than or equal to 0 and less than or equal to 1, and a second calculation step (S5) of inputting the sensor data and count data normalized in the data normalization step (S4) into a pre-learned artificial neural network model to calculate a torque correction value (TQai) for the theoretical torque value (TQbase).

[0040] In addition, the vehicle mechanical compressor torque calculation method may further include a third calculation step (S6) of calculating the output torque (TQout) of the vehicle mechanical compressor (210) by adding the theoretical torque value (TQbase) calculated in the first calculation step (S2) and the torque correction value (TQai) calculated in the second calculation step (S5).

[0041] The above relationship can be a physical equation as follows.

[0042]

[0043] At this time, TQbase is the theoretical torque value, vs is the volume per unit mass of the refrigerant of the air conditioning device (200), Ps is the suction pressure of the vehicle mechanical compressor (210), Pd is the discharge pressure of the vehicle mechanical compressor (210), is the mass flow rate of the refrigerant of the air conditioning device (200), Nc is the rotational speed per minute of the vehicle mechanical compressor (210), and n is the isentropic index.

[0044] Here, in order to reduce the cost, labor, etc. required for sensor installation, only the discharge pressure (Pd) of the vehicle mechanical compressor (210) is the sensor data measured by the sensor (250), and the volume per unit mass (vs) of the refrigerant of the air conditioning device, the suction pressure (Ps) of the vehicle mechanical compressor, and the mass flow rate ( ) may be a value calculated using the sensor data rather than a value measured from the sensor (250). That is, at least one of the data other than the discharge pressure (Pd) of the vehicle mechanical compressor (210) may be obtained by a regression calculation formula combining measured values ​​held without a separate sensor. At this time, it may be preferable that the suction pressure (Ps) of the vehicle mechanical compressor is obtained by calculation. In addition, in order to minimize the cost, labor, etc., it may be preferable that all of the data other than the discharge pressure (Pd) of the vehicle mechanical compressor (210) are obtained by calculation.

[0045] In the above data normalization step (S4), for example, the maximum-minimum normalization method is used, but it is not limited to this and other methods such as a standardization method may be used.

[0046] The above artificial neural network model learns the difference between the theoretical torque value (TQbase) obtained through many tests in advance and the actual measured value (TQreal) of the output torque (TQout) of the vehicle mechanical compressor (210) and establishes its own calculation system, and uses this to calculate the torque correction value (TQai) that is close to the difference between the theoretical torque value (TQbase) and the actual measured value (TQreal). As shown in FIG. 2, the artificial neural network model of the present embodiment can calculate four primary parameters (P1) by combining the sensor data and count data normalized in the data normalization step (S4), calculate three secondary parameters (P2) by combining the four primary parameters (P1), and calculate the torque correction value (TQai) using the three secondary parameters (P2).

[0047] Here, the number and configuration of the primary parameters (P1) and the secondary parameters (P2) may be appropriately modified, but the four primary parameters (P1) and the three secondary parameters (P2) are selected through the artificial neural network model, and an explanation of how the four primary parameters (P1) are obtained by combining the sensor data and the count data, how the three secondary parameters (P2) are obtained by combining the four primary parameters (P1), and how the torque correction value (TQai) is obtained by combining the three secondary parameters (P2) will be omitted.

[0048] However, the more input values ​​considered in the primary parameter (P1), the higher the accuracy of the torque correction value (TQai), and considering the complexity of torque prediction, it may be desirable for the primary parameter (P1) to consider at least four input variables. That is, it may be desirable for the primary parameter (P1) to consider not only the sensor data and the count data, but also various calculation values ​​(e.g., the suction pressure of the vehicle mechanical compressor, the current value applied to the valve (219), the temperature of the second heat exchanger (240), the outside temperature, etc.) that can be obtained using the sensor data and the regression calculation formula. In this case, the suction pressure of the vehicle mechanical compressor and the differential pressure between the discharge pressure and the suction pressure may be calculated based on the sensor data, which is the discharge pressure of the vehicle mechanical compressor, and then used as the input value. Alternatively, based on the discharge pressure of the vehicle mechanical compressor, which is the sensor data, the difference between the outside temperature and the indoor temperature, the difference between the outside temperature and the evaporator temperature, and the difference between the outside temperature and the duct discharge temperature can be calculated and then used as the input value.

[0049] Hereinafter, the operation and effect of the air conditioning device (200) according to the present embodiment will be described.

[0050] That is, when the vehicle is started, the engine (100) rotates, and when the air conditioning device (200) is started, the vehicle mechanical compressor (210) can receive some of the power generated by the engine (100) to suck and compress the refrigerant. That is, when the engine (100) is rotated, the pulley (213) is rotated by being linked to the engine (100) through, for example, a belt, and the rotation shaft (212) can be rotated by contacting the pulley (213) by the clutch (214). Then, the swash plate (215) is rotated together with the rotation shaft (212), the piston (216) reciprocates, and the refrigerant in the suction chamber (V1) is sucked into the compression chamber of the bore (211a), compressed, and then discharged to the discharge chamber (V3). The refrigerant discharged into the discharge chamber (V3) of the vehicle mechanical compressor (210) sequentially passes through the first heat exchanger (220), the expansion unit (230), and the second heat exchanger (240) through a pipe, and then flows back into the suction chamber (V1) of the vehicle mechanical compressor (210). The heat released or absorbed by the refrigerant as it passes through the first heat exchanger (220) and the second heat exchanger (240) can provide cooling and heating to the passenger compartment of the vehicle.

[0051] Here, the air conditioning device (200) according to the present embodiment and the vehicle including the same can be controlled as follows for optimal driving.

[0052] That is, as illustrated in FIG. 3, the sensor (250) can measure the discharge pressure of the mechanical compressor (210) and transmit the sensor data to the air conditioning control unit (270).

[0053] The air conditioning control unit (270) can calculate the output torque (TQout) of the vehicle mechanical compressor (210) based on the sensor data received from the sensor (250) and the count data generated by itself.

[0054] Specifically, referring to FIG. 1, the air conditioning control unit (270) can calculate the theoretical torque value (TQbase) by substituting the sensor data received from the sensor (250) into the relational expression according to the sensor data receiving step (S1) and the first calculation step (S2).

[0055] And, referring to FIGS. 1 and 2, the air conditioning control unit (270) can calculate the torque correction value (TQai) by normalizing the sensor data received from the sensor (250) and the count data generated by itself to a value greater than or equal to 0 and less than or equal to 1 according to the sensor data receiving step (S1), the count data generating step (S3), the data normalization step (S4), and the second calculation step (S5), and then inputting the normalized sensor data and the count data into the artificial neural network model.

[0056] And, referring to FIG. 1, the air conditioning control unit (270) can calculate the output torque (TQout) by adding the torque correction value (TQai) to the theoretical torque value (TQbase).

[0057] And, as shown in FIG. 3, the air conditioning control unit (270) can control the stroke of the vehicle mechanical compressor (210) by controlling the valve (219) of the vehicle mechanical compressor (210) based on the calculated output torque (TQout).

[0058] Specifically, when the output torque (TQout) is large, i.e., when a decrease in the refrigerant discharge amount is required, the valve (219) can be controlled by the air conditioning control unit (270) to increase the amount of refrigerant flowing into the inflow path (not shown) from the discharge chamber (V3). Then, the pressure of the crank chamber (V4) increases, and the sum of the moment of the swash plate (215) due to the pressure of the crank chamber (V4) and the moment due to the return spring of the swash plate (215) (hereinafter, the first moment) can increase. In addition, since the first moment becomes greater than the moment due to the compression reaction force of the piston (216) (hereinafter, the second moment), the inclination angle of the swash plate (215) decreases, the stroke of the piston (216) decreases, and the refrigerant discharge amount can decrease.

[0059] Here, the inclination angle of the swash plate (215) can be measured as the angle between the rotation axis (212) of the swash plate (215) and the normal line of the swash plate (215) based on the rotation center of the swash plate (215).

[0060] On the other hand, when the output torque (TQout) is small, i.e., when an increase in the refrigerant discharge amount is required, the valve (219) can be controlled by the air conditioning control unit (270) to reduce the amount of refrigerant flowing into the inlet passage (not shown) from the discharge chamber (V3). Then, the pressure of the crank chamber (V4) can be reduced, and the first moment can be reduced. And, since the first moment becomes smaller than the second moment, the inclination angle of the swash plate (215) increases, the stroke of the piston (216) increases, and the refrigerant discharge amount can be increased.

[0061] Here, in order to reduce the pressure in the crankcase (V4), the opening amount of the inlet passage (not shown) must be reduced so that the amount of refrigerant flowing into the crankcase (V4) from the discharge chamber (V3) must be reduced, and the refrigerant in the crankcase (V4) must be discharged to the outside of the crankcase (V4). To this end, the discharge passage (218) that guides the refrigerant in the crankcase (V4) to the suction chamber (V1) and the orifice hole that depressurizes the refrigerant passing through the discharge passage (218) to prevent an increase in the pressure in the suction chamber (V1) are provided.

[0062] Meanwhile, when the output torque (TQout) is at an appropriate level, i.e., when it is necessary to maintain the refrigerant discharge amount, the valve (219) can be controlled by the air conditioning control unit (270) to maintain the amount of refrigerant flowing into the inflow path (not shown) from the discharge chamber (V3). Then, the first moment and the second moment are maintained in the same state, the inclination angle of the swash plate (215) is maintained in a steady state, and the stroke of the piston (216) and the refrigerant discharge amount can be maintained constant.

[0063] Here, since the compression reaction force of the piston (216) is proportional to the compression amount, the compression reaction force of the piston (216) and the second moment increase as the inclination angle of the swash plate (215) increases. Accordingly, as the inclination angle of the swash plate (215) increases, the pressure of the crank chamber (V4) for maintaining the inclination angle of the swash plate (215) also increases. That is, when the inclination angle of the swash plate (215) is maintained in a steady state in a relatively large state, a greater pressure is required in the crank chamber (V4) than in the crank chamber (V4) when the inclination angle of the swash plate (215) is maintained in a steady state in a relatively small state.

[0064] In this way, as the stroke of the vehicle mechanical compressor (210) is controlled according to conditions, the vehicle can be operated while providing optimal cooling and heating.

[0065] In this process, in the present embodiment, since the output torque (TQout) is calculated by adding the theoretical torque value (TQbase) obtained through the above relational expression and the torque correction value (TQai) obtained through the artificial neural network model, the output torque (TQout) can be accurately calculated as shown in FIGS. 5 and 6. In addition, since the output torque (TQout) is accurately calculated, the refrigerant discharge amount of the vehicle mechanical compressor (210) is appropriately controlled, so that problems such as a decrease in cooling performance and efficiency, a decrease in vehicle fuel efficiency, and engine stalling can be suppressed.

[0066] Here, unlike the present embodiment, it is possible to consider directly calculating the output torque (TQout) using the artificial neural network model. However, in an actual vehicle environment, there are various and complex conditions such as indoor temperature, outdoor temperature, vehicle condition, vehicle surroundings, and driver operation, and it is impossible to secure data on these conditions. Therefore, it is difficult to develop an artificial neural network model that directly calculates the output torque (TQout). In addition, even if an artificial neural network model that directly calculates the output torque (TQout) is developed, the artificial neural network model is developed only for a specific air conditioning unit (200) and a specific vehicle. Therefore, if conditions change, such as when a component of the air conditioning unit (200) is changed or a vehicle equipped with the air conditioning unit (200) is changed, it takes a considerable amount of time and money to develop the artificial neural network model again. Therefore, it is inefficient to develop an artificial neural network model that directly calculates the output torque (TQout). On the other hand, the artificial neural network model that calculates the torque compensation value (TQai) is relatively simpler than the artificial neural network model that directly calculates the output torque (TQout), and thus is easy to develop and can be easily redeveloped according to changes in conditions. Accordingly, the present embodiment, which calculates the basic theoretical torque value (TQbase) using a predetermined relational expression, calculates the torque compensation value (TQai) to supplement the theoretical torque value (TQbase) using an artificial neural network model, and calculates the output torque (TQout) by adding the torque compensation value (TQai) to the theoretical torque value (TQbase), can be an efficient, effective, universal, and realistic alternative for accurately calculating the output torque (TQout).

[0067] In addition, the present embodiment, by considering the count data as well as the sensor data, reflects the time series data characteristics that change over time after the air conditioning device (200) is started (for example, a change in torque that gradually decreases as the cooling load decreases over time) into the artificial neural network model, thereby enabling more accurate calculation of the torque correction value (TQai). Here, the initial startup time corresponds to the time when the torque is the highest, and in order to identify the initial startup time and increase the prediction accuracy when the torque is high, it may be desirable for the count data to be reset every time the air conditioning device is turned on.

[0068] Meanwhile, the sensor data received from the sensor (250), the count data generated by the air conditioning control unit (270), and the calculated values ​​obtained by calculating from the sensor data may have different scales. However, when calculating the torque correction value (TQai) directly from data having different scales in this way, the dependency of the torque correction value (TQai) on each data becomes proportional to the scale of the data. In other words, the influence of important but small-scale data on the torque correction value (TQai) becomes small, and the influence of unimportant but large-scale data on the torque correction value (TQai) becomes large. Accordingly, a torque correction value (TQai) with low accuracy may be calculated. In consideration of this, the present embodiment normalizes the sensor data received from the sensor (250), the count data generated from the air conditioning control unit (270), and the calculated values ​​obtained by calculating from the sensor data to a value greater than or equal to 0 and less than or equal to 1, and then calculates the torque correction value (TQai) from the normalized data, thereby calculating a torque correction value (TQai) with high accuracy.

[0069] Meanwhile, in the present embodiment, the above-mentioned theoretical torque value (TQbase) is calculated using the above-mentioned relationship based on a physical formula, but it is not limited thereto. That is, after deriving an empirical formula from data acquired through multiple experiments in advance, the theoretical torque value (TQbase) can also be calculated using the derived empirical formula. Of course, in this case, the artificial neural network model should be trained to calculate a torque correction value (TQai) close to the difference between the theoretical torque value (TQbase) calculated using the empirical formula and the actual value (TQreal) of the output torque (TQout).

Claims

1. A sensor data receiving step for receiving sensor data measured from a sensor installed in an air conditioning device including a mechanical compressor for a vehicle; A first calculation step of calculating a theoretical torque value of the vehicle mechanical compressor from the sensor data through a predetermined relational expression; A second calculation step of calculating a torque correction value for the theoretical torque value from the sensor data using a pre-learned artificial neural network model; A method for calculating torque of a mechanical compressor for a vehicle, comprising a third calculation step of calculating the output torque of the mechanical compressor for a vehicle by adding the theoretical torque value and the torque correction value.

2. In paragraph 1, Further comprising a count data generation step for generating count data for the operating time of the mechanical compressor for the vehicle, The above artificial neural network model is a method for calculating the torque of a mechanical compressor for a vehicle by calculating the torque compensation value from the sensor data and the count data.

3. In paragraph 2, A method for calculating the torque of a mechanical compressor for a vehicle, wherein the above count data is reset each time the air conditioning device is turned on.

4. In paragraph 2, Further comprising a data normalization step of normalizing the sensor data and the count data to a number greater than or equal to 0 and less than or equal to 1, The above artificial neural network model is a method for calculating the torque compensation value of a mechanical compressor for a vehicle from normalized sensor data and count data.

5. In paragraph 2, A method for calculating torque of a mechanical compressor for a vehicle, wherein the artificial neural network model calculates a plurality of primary parameter values ​​from the sensor data and the count data, calculates at least one secondary parameter value by combining the plurality of primary parameter values, and calculates the torque correction value using the at least one secondary parameter value.

6. In paragraph 5, A method for calculating torque of a mechanical compressor for a vehicle, characterized in that there are at least four input variables input to the artificial neural network model.

7. In paragraph 1, The above relationship is, A method for calculating torque of a mechanical compressor for a vehicle, wherein TQbase is the theoretical torque value, vs is the volume per unit mass of the refrigerant of the air conditioning device, Ps is the suction pressure of the mechanical compressor for a vehicle, Pd is the discharge pressure of the mechanical compressor for a vehicle, is the mass flow rate of the refrigerant of the air conditioning device, Nc is the rotational speed per minute of the mechanical compressor for a vehicle, and n is the eigentropic index.

8. In paragraph 7, A method for calculating torque of a mechanical compressor for a vehicle, characterized in that the discharge pressure of the mechanical compressor for a vehicle is sensor data measured by the sensor.

9. In paragraph 7, A method for calculating torque of a mechanical compressor for a vehicle, characterized in that at least one of the volume per unit mass of the refrigerant of the air conditioning device, the suction pressure of the mechanical compressor for a vehicle, and the mass flow rate of the refrigerant of the air conditioning device is a value calculated using the sensor data.

10. In paragraph 9, A method for calculating torque of a mechanical compressor for a vehicle, characterized in that the suction pressure of the mechanical compressor for a vehicle is a value calculated using the sensor data.

Images