Method for estimating a required oil volume flow in a work machine, method for training a machine learning model, computer program, device and work machine with a device

Abstract

The present invention relates to a method for estimating a required oil volume flow (119) in a work machine (100). The method comprises obtaining (S1) an oil pressure value (112), obtaining (S2) a value representative of a current oil volume flow (118), obtaining (S3) a value of an excess oil volume flow (116), inputting (S4) the oil pressure value (112), the value representative of the current oil volume flow (118), and the value of the excess oil volume flow (116) as input signals into a trained machine learning model (110), and determining (S5) an estimated value of the required oil volume flow (119) as an output of the trained machine learning model (110) based on the output data of the trained machine learning model (110).The invention also relates to a method for training a machine learning model (110), a computer program, a device (109) and a work machine (100) with a device (109).

Description

Technical area

[0001] The present invention relates to a method for estimating a required oil volume flow in a work machine, a method for training a machine learning model, a computer program, a device and a work machine with a device. State of the art

[0002] In the state of the art, a maximum possible oil volume flow is pumped in a working machine.

[0003] DE 10 2021 211 967 A1 discloses a hydraulic machine with an adjustable displacement volume for the pressure-controlled pressure medium supply to at least one hydrostatic consumer.

[0004] DE 10 2015 225 649 A1 presents a method for controlling and / or regulating a hydraulic system of a transmission, wherein hydraulic oil is pumped by a pump.

[0005] DE 10 2013 217 708 A1 relates to a method for operating a drive system having a hydrostatic drive unit. Description of the invention

[0006] The present invention relates in a first aspect to a method for estimating a required oil volume flow in a working machine.

[0007] The work machine may comprise a machine with a common oil supply. A common oil supply may mean that the work machine may comprise hydraulic systems, transmissions, and engines, and these may use a common oil tank. A work machine may be a construction machine, such as a hydraulic excavator or a dump truck, an agricultural machine, such as a tractor or combine harvester, or a commercial vehicle, such as a bus. The work machine may comprise a vehicle, such as a motor vehicle or a truck.

[0008] The required oil flow rate can be the oil flow rate that a pump delivers per unit of time, which is adjusted to the actual oil demand in the working machine. The required oil flow rate can be the difference between the actual oil flow rate and the excess oil flow rate.

[0009] Estimating can include determining or calculating.

[0010] The method may be a computer-implemented method. The steps of the method may be executed by a computer using a corresponding computer program product, wherein the computer may be a control unit or part of a control unit.

[0011] The method includes obtaining an oil pressure value, obtaining a value representative of a current oil volume flow, and obtaining a value of an excess oil volume flow.

[0012] Obtaining a value may include obtaining a raw sensor signal. Obtaining may include obtaining a processed sensor value. Obtaining may also include reading the value, for example, from a database. Obtaining the value may occur via a CAN, for example, a CAN bus.

[0013] The oil pressure value can include a lubricant pressure value. The oil pressure value can correlate with the oil flow rate and thus with an excess oil flow rate. The oil pressure value can include a pressure differential of a pump in a load-sensing system. The pressure differential can include a difference between a load pressure and a pump pressure.

[0014] The value representative of the current oil volume flow can include a value of the current oil volume flow. The current oil volume flow can be a pumped oil volume flow of the pump, for example a fixed displacement pump. The value representative of the current oil volume flow can include a value that is indicative of the current oil volume flow. The value representative of the current oil volume flow can make it possible to draw conclusions about the current oil volume flow. The value representative of the current oil volume flow can be determined based on a pump characteristic curve. The value representative of the current oil volume flow can be determined based on a flow measurement. The value representative of the current oil volume flow can be determined based on a performance calculation.

[0015] The excess oil flow rate can be the oil flow rate pumped by a charge pump per unit of time in excess of the required amount. The excess oil flow rate can be an excess of the oil flow rate at a lubrication pressure valve. The excess oil flow rate can be determined using an estimation method.

[0016] The method comprises inputting the oil pressure value, the value representative of the current oil volume flow, and the value of the excess oil volume flow as input signals into a trained machine learning model, wherein the machine learning model was trained based on training data each having an oil pressure value, a value representative of a current oil volume flow, and a value of an excess oil volume flow as input data, and a value of a required oil volume flow as output data.

[0017] The training data can be generated, for example, as part of a test setup. For this purpose, a volumetric flow meter can be integrated at a location in the working machine. The volumetric flow meter can be arranged in a hydraulic consumer of the working machine, for example in a gearbox. The volumetric flow meter can be arranged in a line of the oil pressure valve. The oil pressure valve can be a passive pressure relief valve. Furthermore, an oversized constant-displacement pump can be used to acquire the training data of the machine learning model. The volumetric flow meter can thus measure an excess oil volume flow in the working machine. The additional training data can also be measured. For example, the oil pressure value can comprise a measured oil pressure value.The value representative of the current oil flow rate from the training data can be measured using a volumetric flow meter. The value representative of the current oil flow rate from the training data can be determined based on the pump characteristic curve or the power calculation. The required oil flow rate can be calculated from the difference between the measured oil flow rate and the measured excess oil flow rate and used as the input data for the machine learning model. Thus, the machine learning model can be trained based on the measured and / or determined values. The training data can also be provided partially or completely via a simulation.

[0018] The training data can include measurements across a variety of or all relevant states of the working machine. The training data can thus represent a representative representation of the relevant states of the working machine and load spectrums. Mapping a variety of or all relevant states of the working machine can, for example, enable a more accurate estimation of the required oil flow rate.

[0019] Before entering the training data to train the machine learning model, the training data can be prepared.

[0020] The method further comprises determining an estimate of the required oil volume flow as an output of the machine learning model based on the output data of the trained machine learning model.

[0021] The estimated value of the required oil flow rate may be the output of the trained machine learning model. The estimated value of the required oil flow rate may include post-processing the output of the trained machine learning model. The post-processing may, for example, include using the determined estimated value of the required oil flow rate as a control input.

[0022] The method can provide an estimate of the required oil flow rate. This allows fluctuations in the oil demand of the working machine to be recorded. In particular, these fluctuations can be detected early on because the machine learning model was trained with training data representative of as many working machine states as possible. Thus, a time delay in the required oil flow rate can be recorded. This can enable changes in parameters that lead to a change in oil demand to be identified early on by the machine learning model. Furthermore, the required oil flow rate can be optimized based on specific properties of the working machine, for example, a lubricant used, a lubrication system used, and / or a size of the working machine.Furthermore, this method can be used to optimally adjust the required oil volume flow based on specific conditions of the working machine, such as vibrations, operating time and / or temperatures.

[0023] In one embodiment, the method may comprise outputting the required oil volume flow based on the determined estimate of the required oil volume flow, wherein the output is suitable to be used as a control input for a controller.

[0024] Outputting may include sending the required oil volume flow, for example, as an output signal. Outputting may include displaying the required oil volume flow, for example, on a user display. The required oil volume flow may be output in the form of the determined estimated value of the required oil volume flow. Outputting the required oil volume flow may include processing the determined estimated value of the required oil volume flow before outputting.

[0025] The output of the required oil flow rate can be used as a control input for the controller, for example, a pump controller. The method can thus include controlling a controller based on the output of the required oil flow rate.

[0026] Control can be a dynamic process involving continuous measurement and adjustment. This can, for example, involve comparing an actual value with a target value. Control can also be the setting of a specific variable that is adjusted and occurs independently of continuous measurements or adjustments.

[0027] This embodiment allows a controller to be controlled according to the actual oil requirement. This can reduce the pump's energy consumption because reactive power can be reduced. This allows the pump's energy requirement to be optimized. Furthermore, the method can provide a time for control or pre-control before the actual oil requirement increases. The method can therefore provide efficient control for a controller. For example, the control for the pump can be set so that an adjusted oil volume flow is pumped at all times to ensure the oil requirement is met. The method can therefore provide a control input for a demand-based control strategy. In general, the output can enable further processing of the output required oil volume flow.

[0028] In a further embodiment, if a threshold value of the output required oil volume flow as a control input is exceeded, the use of the output of the required oil volume flow as a control input can be suppressed.

[0029] The control input can have a permissible range that defines the permissible range of the required oil volume flow output. An excessively high value for the required oil volume flow output as the control input can be interpreted as exceeding a threshold. An excessively low value for the required oil volume flow output as the control input can also be interpreted as exceeding a threshold.

[0030] If the threshold value of the required oil volume flow is exceeded as a control input, the method can suppress the use of the output of the required oil volume flow as a control input. Suppressing the use can comprise not using the output required oil volume flow as a control input. Suppressing the use can mean that a specified value is used as the control input instead of the output required oil volume flow. This specified value can lie within a permissible control range. Suppressing the use can comprise setting a control input to a maximum permissible control input value. Suppressing the use can comprise setting a control input to a minimum permissible control input value. The use of the maximum or minimum permissible control input value can depend on whether the output of the required oil volume flow was too large or too small.

[0031] This allows for threshold-based control. This ensures that the required oil flow output as the control input lies within a definable permissible range. This increases safety, as only permissible values ​​are possible as the control input.

[0032] In a further embodiment, if a change rate of the control input is exceeded, use of the output of the required oil volume flow as a control input can be suppressed.

[0033] The rate of change of the control input can have a permissible range, so that if the rate of change is exceeded, the use of the required oil flow output as a control input is suppressed. The rate of change can be a change in the output required oil flow per unit of time.

[0034] Suppressing usage may include not using the required oil volume flow output as a control input. Instead of the required oil volume flow output, a specified value may be used as the control input. This specified value may be within a permissible control range. Suppressing usage may include setting a control input to a maximum permissible control input value. Furthermore, suppressing usage may include using a value with a permissible rate of change. The permissible rate of change may be a permissible deviation from a previous value. Suppressing usage may include setting a control input to a minimum permissible control input value.

[0035] This allows for gradient-based control. This prevents sudden changes in the required oil flow rate output as the control input. This increases safety, as only permissible values ​​are possible as the control input.

[0036] In a further embodiment, the method may include a step of obtaining a further value. In the inputting step, the obtained further value may be input into the trained machine learning model as an input signal, wherein the training data comprises a further value as input data, wherein the further value comprises one or more of a value of a hydraulic system, a value of a priority system, a value of a state of an actuator, a transmission variable, or an oil temperature value.

[0037] The hydraulic system may comprise at least one hydraulic circuit. The hydraulic circuit may comprise hydraulic valves. The hydraulic system may comprise a control system, for example, a steering system or brakes, hydraulics for operating attachments, for example, a bucket or arm of an excavator, and a travel drive. Different hydraulic systems may have different oil requirements. When a hydraulic system is switched on or off, the required oil flow rate may change depending on the oil demand of the hydraulic system. This change can be accounted for by considering the value of the hydraulic system. Consequently, the value of the hydraulic system may be relevant for estimating the required oil flow rate.

[0038] The priority system can include a value that describes the importance of a hydraulic function or hydraulic system. For example, a travel drive may have a higher priority than the use of an attachment. When using the attachment, it may be important that it continues to receive a sufficiently high oil flow to ensure fault-free operation. For example, connecting another hydraulic system with a high priority can result in a rapid increase in the required oil demand. Thus, the value of the priority system can be relevant for estimating the required oil flow.

[0039] The hydraulic valve may be part of the load pressure sensing system. The hydraulic valve may comprise an actuator. The actuator may comprise a binary actuator. A binary actuator may have only two states, for example, on / off or active / inactive. The actuator may comprise a proportional or variable actuator. This actuator may comprise intermediate states or ranges of states. This may enable targeted control.

[0040] The state of an actuator can thus comprise an operating state. The state of an actuator can comprise an operating position. The operating state can comprise a state of an actuator, for example, on / off, active / inactive, or a fault state. The operating position can comprise an actuator position. The position of an actuator can be indicative of an intermediate state. The position of an actuator can comprise a required oil pressure value to move the actuator to a specific position. The state of the actuator can therefore influence the oil requirement. Thus, the value of the actuator state can be relevant for estimating the required oil flow rate.

[0041] The gearbox size can be indicative of the load on the gearbox. The gearbox size can, for example, include the gearbox speed. The gearbox speed can include an input speed or an output speed. The gearbox speed can, for example, influence heat generation within the gearbox and consequently the oil demand. Thus, the gearbox size can be relevant for estimating the required oil flow rate.

[0042] For example, the oil temperature value can influence the viscosity of the oil. Thus, the oil temperature value can be relevant for estimating the required oil flow rate.

[0043] The additional value can also include a pump speed.

[0044] Obtaining the further value may include obtaining a raw signal from a sensor. Obtaining may include obtaining a processed sensor value. Obtaining may also include reading the value, for example, from a database. Obtaining the value may occur via a CAN, for example, a CAN bus. The manner of obtaining the further value may depend on the further value. For example, the transmission size may be obtained via a transmission CAN. The value of the hydraulic system, the value of the priority system, and the value of the state of an actuator may be obtained, for example, via a vehicle CAN.

[0045] This allows additional values ​​to be received to determine the current working machine status. These additional values ​​can then be taken into account when estimating the required oil flow rate. Consequently, an accurate estimate of the required oil flow rate is possible depending on the working machine's status.

[0046] Furthermore, this allows for early detection when a change in the working machine's condition causes a change in the required oil flow. This can ensure sufficient time for pre-regulation. This can increase safety and ensure fault-free functionality.

[0047] In another embodiment, the machine learning model may be a neural network.

[0048] However, it is also possible that the machine learning model is a Support Vector Regression (SVR) model, or a Gaussian Process Regression (GPR) model, or a Convolutional Neural Network (CNN).

[0049] The neural network can be a regression model. The neural network can be a fully connected neural network (FCNN). This allows the required computing power to be kept low. Thus, the method can be executed on common control units of work machines.

[0050] Using a neural network can provide a continuous value mapping. The continuous value mapping can be based on non-discrete classes. The neural network can also have a classification with discrete classes as output.

[0051] In another embodiment, the machine learning model may be a decision tree model.

[0052] The decision tree model can have a staged output. The output stages can be defined based on thresholds. A threshold of the decision tree model can be set depending on the total oil volume in the working machine. The stages can also be defined based on the working machine's operating points. The stages can be selected so that the oil demand changes significantly when transitioning between the stages.

[0053] A second aspect relates to a method for training a machine learning model for use in estimating a required oil volume flow in a method according to an embodiment of the first aspect. Training the machine learning model comprises the steps of inputting training data to the machine learning model, wherein the training data each comprise an oil pressure value, a value representative of a current oil volume flow, and a value of an excess oil volume flow as input data, and a value of a required oil volume flow as output data.

[0054] This training data can, for example, result from an experimental setup described above. The aforementioned additional input of another value as input data can also be possible in the training process. This additional value can also include the speed of the fixed-displacement pump. The process for training the machine learning model can include supervised learning, also known as a supervised learning method.

[0055] A third aspect relates to a computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the steps of the method for estimating an excess oil volume flow in a working machine.

[0056] A fourth aspect relates to a device configured to perform steps of a method according to an embodiment of the first aspect or the second aspect.

[0057] A fifth aspect relates to a work machine with a device for carrying out a method according to an embodiment of the first aspect, wherein the work machine comprises a pump, an oil tank and a hydraulic consumer.

[0058] The hydraulic consumer can be one or more of a gearbox, a motor, a braking system, a cylinder, a pump, a steering system, a valve or similar.

[0059] Advantages and features described with regard to the method also apply to the device, the work machine, and the computer program, and vice versa. These are therefore described only once. Short description of the characters Fig. 1 shows schematically a work machine with a device according to an embodiment. Fig. Figure 2 schematically shows a trained machine learning model in a usage state. Fig. 3 shows a flowchart of a method for estimating an excess oil volume flow in a working machine. Detailed description of embodiments

[0060] Fig. 1 schematically shows a work machine 100 with a device 109 according to one embodiment.

[0061] The work machine 100 comprises an oil pressure sensor 102, an oil temperature sensor 104, an oil pressure valve 105, a pump 106, an oil tank 108, and a hydraulic consumer 107 in the form of a working hydraulic system 107a. The pump 106 is part of a load pressure reporting system. The work machine 100 further comprises a device 109 with a trained machine learning model 110. The device 109 is designed as a control unit. The device 109 is communicatively connected to the oil pressure sensor 102, the oil temperature sensor 104, the oil pressure valve 105, the working hydraulic system 107a, and the pump 106. Furthermore, the pump 106, the oil tank 108, the oil pressure valve 105, and the working hydraulic system 107a are in fluid communication. During operation, pump 106 delivers a volume flow of oil from oil tank 108 to supply lubricant to hydraulic consumer 107 in the form of working hydraulics 107a. Working hydraulics 107a includes an actuator.The actuator is designed as a proportional actuator and can have two final states as well as further intermediate states to perform a work function.

[0062] The machine learning model 110 is a Fully Connected Neural Network (FCNN).

[0063] The oil temperature sensor 104 detects an oil temperature value 114. The oil pressure sensor 102 detects an oil pressure value 112 in the form of a pressure differential. The pressure differential is the difference between a load pressure and a pump pressure. The pressure differential can also be used to determine a value representative of the current oil volume flow 118.

[0064] Device 109 receives the oil pressure value 112, the oil temperature value 114, and the value representative of the current oil volume flow 118. Device 109 also receives a value 117 of a state of the proportional actuator of the working hydraulics 107a. This value describes an operating position of the actuator. The obtained value is indicative of one of the end positions or one of the intermediate positions of the actuator. Furthermore, device 109 receives a value of an excess oil volume flow 116.

[0065] These obtained values ​​are input by the device 109 as an input signal to the trained machine learning model 110. Based on the output data of the trained machine learning model 110, an estimated value of the required oil volume flow 119 is determined.

[0066] The determined estimated value of the required oil volume flow 119 is then used to control the pump 106 of the working machine 100 in order to adjust the oil volume flow delivered by the pump 106 according to a state of the working machine 100. As a result, the performance of the pump 106 is efficiently controlled based on the determined estimated value of the required oil volume flow 119.

[0067] In Fig. 2 shows a device 109 in the form of a control unit with a trained machine learning model 110.

[0068] The trained machine learning model 110 receives an oil pressure value 112, an oil temperature value 114, a value of an excess oil volume flow 116, a value of an actuator state 117, and a value representative of a current oil volume flow 118. The machine learning model 110 was trained with correspondingly measured input data. The training data includes a value of the required oil volume flow 119 of the work machine 100 for training the output of the machine learning model 110. Accordingly, the trained machine learning model 110 determines an estimated value of the required oil volume flow 119 as output.

[0069] Fig. 3 shows a flowchart of a method for estimating a required oil volume flow 119 in a work machine 100.

[0070] The method comprises a step S1 of obtaining an oil pressure value 112. A further step S2 comprises obtaining a value representative of a current oil volume flow 118. Step S3 comprises obtaining a value of an excess oil volume flow 116. These values ​​are input to the trained machine learning model 110 as input signals in a step S4. In a step S5, an estimated value of the required oil volume flow 119 is determined based on the output data of the trained machine learning model 110.

[0071] In an additional step S6, the required oil volume flow is output based on the determined estimated value of the excess oil volume flow 119. In a further step S7, the output of the required oil volume flow is used as a control input for controlling a pump 106 of the work machine 100. Reference symbol 100 work machines 102 Oil pressure sensor 104 Oil temperature sensor 105 Oil pressure valve 106 Pump 107 hydraulic consumer 107a Working hydraulics 108 Oil tank 109 Device 110 Machine Learning Model 112 Oil pressure value 114 Oil temperature value 116 Value of an excess oil volume flow 117 Value of a state of an actuator 118 Value representative of a current oil volume flow 119 Value of the required oil volume flow S1 Obtaining an oil pressure value S2 Obtaining a value representative of a current oil volume flow S3 Obtaining a value of an excess oil volume flow S4 Input into a trained machine learning model S5 Determine an estimate of the required oil flow rate S6 Output of the required oil volume flow S7 Control of a controller

Claims

[1] Method for estimating a required oil volume flow (119) in a working machine (100), comprising the steps: - Obtaining (S1) an oil pressure value (112); - Obtaining (S2) a value representative of a current oil volume flow (118); - Obtaining (S3) a value of an excess oil volume flow (116); - inputting (S4) the oil pressure value (112), the value representative of the current oil volume flow (118) and the value of the excess oil volume flow (116) as input signals into a trained machine learning model (110), wherein the machine learning model (110) was trained based on training data which each have an oil pressure value (112), a value representative of a current oil volume flow (118) and a value of an excess oil volume flow (116) as input data and a value of a required oil volume flow (119) as output data; - Determining (S5) an estimated value of the required oil volume flow (119) as output of the trained machine learning model (110) based on the output data of the trained machine learning model (110). [2] Method according to claim 1, characterized by in that the method comprises a step of outputting (S6) the required oil volume flow (119) based on the determined estimated value of the required oil volume flow (119), wherein the output is suitable for being used as a control input for a controller (106). [3] Method according to claim 2, characterized by that when a threshold value of the output required oil volume flow (119) as a control input is exceeded, use of the output of the required oil volume flow (119) as a control input is suppressed. [4] Method according to one of claims 2 or 3, characterized bythat if a change rate of the control input is exceeded, the use of the output of the required oil volume flow (119) as a control input is suppressed. [5] Method according to one of the preceding claims, characterized by in that the method comprises a step of obtaining a further value and, in the step of inputting, the obtained further value is input as an input signal into the trained machine learning model (110), wherein the training data comprise a further value as input data, wherein the further value comprises one or more of a value of a hydraulic system, a value of a priority system, a value of a state of an actuator (117), a transmission variable or an oil temperature value (114). [6] Method according to one of the preceding claims, characterized bythat the machine learning model (110) is a neural network, or a support vector regression (SVR) model, or a Gaussian process regression (GPR) model, or a convolutional neural network (CNN). [7] Method according to one of claims 1 to 5, characterized by that the machine learning model (110) is a decision tree model. [8] A method for training a machine learning model (110) for use in estimating a required oil volume flow (119) in a method according to any one of the preceding claims, wherein the training comprises the steps of: Inputting training data into the machine learning model (110), wherein the training data each have an oil pressure value (112), a value representative of a current oil volume flow (118) and a value of an excess oil volume flow (116) as input data and a value of a required oil volume flow (119) as output data. [9] A computer program comprising instructions which, when executed by a computer, cause the computer to carry out the steps of the method according to any one of claims 1 to 7. [10] Apparatus (109) adapted to carry out steps of the method according to any one of claims 1 to 8. [11] Work machine (100) with a device (109) for carrying out a method according to one of claims 1 to 7, wherein the work machine (100) comprises a pump (106), an oil tank (108) and a hydraulic consumer (107, 107a).

Images