Method for determining refrigerant mass in refrigerant circuit of motor vehicle, and motor vehicle

Abstract

The invention relates to a motor vehicle (10) and to a method for determining a refrigerant mass in a refrigerant circuit (12) of a motor vehicle (10), wherein a compressor (14) feeds compressed refrigerant to a refrigerant cooler (16). Refrigerant from the refrigerant cooler (16) is expanded by means of an expansion mechanism (22). The expanded refrigerant is fed to the evaporator (24). Refrigerant from the evaporator (24) is fed to a suction side (28) of the compressor (14). At least one input parameter which is dependent on the operation of the refrigerant circuit (12) is fed to a model (32) which outputs a model parameter which is used to determine the refrigerant mass. The model (32) outputs a refrigerant temperature at a predetermined location of the refrigerant circuit (12) as a model parameter. The model parameter (32) is compared with a measured value of the refrigerant temperature detected at the predetermined point. The refrigerant mass is estimated based on a comparison of the model parameters to the measurements.

Description

Technical Field

[0001] The invention relates to a method for determining the amount of refrigerant in a refrigerant circuit of a motor vehicle, wherein a compressor of the refrigerant circuit delivers compressed refrigerant to at least one refrigerant cooler of the refrigerant circuit. The refrigerant from the refrigerant cooler is expanded / decompressed by means of at least one expansion mechanism of the refrigerant circuit. The expanded refrigerant is delivered to at least one evaporator of the refrigerant circuit. The refrigerant from at least one evaporator is delivered to the suction side of the compressor. At least one input parameter related to the operation of the refrigerant circuit is delivered to a model, which outputs at least one model parameter. The at least one model parameter is used to determine the amount of refrigerant. In addition, the invention relates to a motor vehicle having a refrigerant circuit. Background Art

[0002] In a motor vehicle, the refrigerant provided for operating a refrigerant circuit is initially set as a lifetime filling quantity in terms of the initial filling quantity or initial refrigerant quantity. Accordingly, the refrigerant quantity initially present in the refrigerant circuit is intended to remain constant over the service life of the refrigerant circuit or the motor vehicle.

[0003] In practice, however, the refrigerant may gradually be lost, i.e. the amount of refrigerant present in the refrigerant circuit decreases compared to the initial filling amount. Such a loss may be noticed by the user of the motor vehicle in the form of a reduction in the refrigeration capacity of the refrigerant circuit. Accordingly, a reduction in refrigeration capacity may occur when the refrigerant circuit is underfilled with refrigerant, i.e. when less refrigerant is present in the refrigerant circuit than was originally set. This is disadvantageous.

[0004] In addition to insufficient cooling capacity, a significantly insufficient refrigerant circuit or refrigerant circuit filling can also lead to the possibility of compressor damage, because if too little refrigerant is available in the refrigerant circuit, it may happen that not enough oil for lubrication is supplied to the compressor together with the refrigerant.

[0005] In order to avoid a significant reduction in the cooling capacity of the refrigerant circuit and the associated complaints from motor vehicle users, for example, fixed maintenance intervals can be set to fill the refrigerant circuit with refrigerant. However, it should be noted that the missing amount of refrigerant cannot be simply replenished in the refrigerant circuit. This is because the actual amount of refrigerant filled in the refrigerant circuit cannot be easily measured.

[0006] Furthermore, it is important for the reliable operation of the refrigerant circuit to maintain a refrigerant quantity that is as accurately defined as possible in the refrigerant circuit or refrigerant cycle. Accordingly, the refrigerant circuit should be filled with neither more nor less than the refrigerant quantity specified for the respective refrigerant circuit of the motor vehicle. In order to always achieve the correct filling of the refrigerant circuit with refrigerant, all the refrigerant must first be removed from the refrigerant circuit during such maintenance. The refrigerant circuit must then be refilled with exactly the refrigerant quantity defined for the refrigerant circuit.

[0007] A disadvantage here is the fact that such regular replacement of the refrigerant is associated with high maintenance costs. Conversely, if such regular maintenance is omitted, this is accompanied by the possibility of a reduction in the refrigeration capacity of the refrigerant circuit. Moreover, if the user of the motor vehicle does not notice or ignores the lack of refrigerant, this may even lead to damage or even possible destruction of the compressor. This is also disadvantageous.

[0008] Furthermore, if the refrigerant medium circuit is not optimally filled with refrigerant medium, this results in inefficient operation of the refrigerant medium circuit.

[0009] Therefore, it is desirable to ascertain or determine the amount of refrigerant present in the refrigerant circuit as accurately as possible, because if the amount of refrigerant actually present in the refrigerant circuit deviates from the amount of refrigerant set for the refrigerant circuit, this deviation can be responded to simply and specifically.

[0010] DE 10 2013 019 498 A1 describes a method for operating an air conditioning system of a motor vehicle, wherein a refrigerant circulates in a refrigerant circuit of the air conditioning system. Filling information of the refrigerant circuit is calculated from state parameters of the refrigerant circuit by means of a data-driven mathematical model. As input data taken into account in the mathematical model for calculating the filling information, compressor power, compressor rotational speed, compressor suction pressure, measured pressure in the high-pressure region of the refrigerant circuit, and temperatures upstream and downstream of the compressor can be used.

[0011] Thus, in this method, the filling state, ie the amount of refrigerant present in the refrigerant circuit, is determined directly based on these state parameters or input parameters by means of a data-driven mathematical model. However, in this method, there is the potential for improving the accuracy of the filling state determination result.

[0012] WO 2007 / 047886 A1 describes a method in which an algorithm generates a refrigerant filling state model based on an observed refrigerant filling state. The algorithm determines an expected refrigerant filling state based on the model and compares the existing refrigerant filling state with the expected refrigerant filling state. The existing refrigerant filling state is determined with the aid of a filling state indicator, which should provide accurate results when the system is running and in a stable state.

[0013] However, such measurement-based detection of the filling level, for example, in a refrigerant reservoir is often subject to significant errors, in particular due to the turbulence of the flowing refrigerant in the refrigerant reservoir during operation of the refrigerant circuit. Summary of the invention

[0014] The object of the present invention is to specify a method of the type mentioned in the introduction, by means of which the amount of refrigerant can be estimated with particularly high accuracy, and to specify a motor vehicle which is designed to carry out the method.

[0015] This object is achieved by a method having the features of claim 1 and by a motor vehicle having the features of claim 10. Advantageous embodiments of the invention with expedient developments are specified in the dependent claims and in the following description.

[0016] In a method for determining the amount of refrigerant in a refrigerant circuit of a motor vehicle according to the present invention, a compressor of the refrigerant circuit delivers compressed refrigerant to at least one refrigerant cooler of the refrigerant circuit. Depending on the refrigerant used, at least one refrigerant cooler can be designed as a gas cooler or a condenser, by means of which the refrigerant flowing through the refrigerant cooler can be caused to condense. The refrigerant from the refrigerant cooler is expanded by means of at least one expansion mechanism of the refrigerant circuit. The expanded refrigerant is delivered to at least one evaporator of the refrigerant circuit. The refrigerant from at least one evaporator is delivered to the suction side of the compressor. At least one input parameter related to the operation of the refrigerant circuit is delivered to a model, which outputs at least one model parameter. The at least one model parameter is used to determine the amount of refrigerant. Here, as at least one model parameter, the model outputs a refrigerant temperature at a predetermined location of the refrigerant circuit. The model parameter, i.e., the refrigerant temperature predicted at a predetermined location of the refrigerant circuit by means of the model, is compared with a measured value of the refrigerant temperature detected at a predetermined location of the refrigerant circuit. The amount of refrigerant is estimated based on the comparison of the model parameter with the measured value.

[0017] The basis for this is the recognition that a direct estimation of the amount of refrigerant in the refrigerant circuit based on operating data of the refrigerant circuit or the refrigerant system (as described in DE 10 2013 019 498 A1) has proven to be relatively difficult or inaccurate. This method is particularly suitable if the refrigerant circuit contains a refrigerant collector, in particular if the refrigerant collector is a relatively large refrigerant collector.

[0018] Because, in a refrigerant circuit with a refrigerant collector, the volume flow of the refrigerant guided in the refrigerant circuit is related to the operating point of the refrigerant circuit. Accordingly, in the case of a low load or a low refrigeration power required and to be provided by the at least one evaporator, less refrigerant is guided in the circulation and more refrigerant is stored in the collector, compared to the case of a high load or a high refrigeration power required of the at least one evaporator. Accordingly, depending on the corresponding operating point or operating state of the refrigerant circuit, only a part of the refrigerant is sucked into the compressor or is delivered to the suction side of the compressor and then compressed.

[0019] The operating data recorded at the compressor therefore do not adequately reflect the presence of refrigerant in the refrigerant receiver. This is particularly the case when the refrigerant circuit is operated at low load. This is because it is very difficult to determine, especially at low load, whether the refrigerant present in the refrigerant circuit is less than the refrigerant required to provide the required refrigeration power to be provided by the at least one evaporator.

[0020] In the method described here, the amount of refrigerant actually present in the refrigerant circuit or the amount of refrigerant is estimated indirectly and based on a model. This is because it has been shown that this indirect estimation of the amount of refrigerant, which is based on a comparison of model parameters with measured values ​​of the refrigerant temperature at at least one predetermined point in the refrigerant circuit, provides better and more accurate results than a direct determination of the amount of refrigerant, in which at least one input parameter is also used. Thus, with the aid of this method, in which the model parameters are compared with the measured values, the amount of refrigerant can be estimated with particularly high accuracy.

[0021] That is, the actual amount of refrigerant is not estimated directly based on the input parameters, but the model parameters are first output based on the input parameters. And the actual amount of refrigerant is estimated only after comparing the model parameters with the measured value of the temperature. Therefore, only the temperature is directly determined as a model parameter based on the input parameters, and then the actual amount of refrigerant is indirectly inferred from the model parameters.

[0022] The method is based on the idea of ​​creating at least one model for the temperature prevailing at at least one predetermined location of the refrigerant circuit, which model reflects the behavior of model parameters at a certain refrigerant quantity. The deviation between the temperature measurement values ​​and the model parameters predicted by the model that occurs during the operation of the refrigerant circuit can then be used to estimate the refrigerant quantity.

[0023] In this regard, it has been shown that, in particular when using correlations, at least one model parameter in the form of a refrigerant temperature at at least one predetermined location of the refrigerant circuit can be used very well to determine the total refrigerant quantity actually present in the refrigerant circuit. Furthermore, since the refrigerant temperature present at the at least one predetermined location can be detected very easily in a measurement-technical manner, the refrigerant quantity can be estimated simply and specifically based on a comparison of the at least one model parameter with the measured value. In this way, reliable information about the actual or current refrigerant filling state of the refrigerant circuit can be obtained during the continuous operation of the refrigerant circuit.

[0024] Accordingly, in the event of a difference between the amount of refrigerant present and the amount of refrigerant provided for the refrigerant circuit, measures can be taken very simply and with little effort to ensure that the amount of refrigerant provided for the refrigerant circuit is adjusted in the refrigerant circuit. Such measures can, for example, include compensating for a lack of refrigerant or, in the event of an excess of refrigerant in the refrigerant circuit, reducing the amount of refrigerant.

[0025] The method is particularly suitable for determining the amount of refrigerant in the refrigerant circuit when the refrigerant circuit includes a refrigerant collector. In such refrigerant circuits with a refrigerant collector, it is usually difficult to determine the amount of refrigerant, especially because the filling state of the refrigerant in the refrigerant collector cannot be simply and directly measured. However, these difficulties are overcome in the following way: the amount of refrigerant is estimated based on the minimum amount present by comparing the model parameters with the measured values.

[0026] Therefore, the refrigerant medium circuit preferably has a refrigerant medium collector, wherein the refrigerant medium collector can be arranged on the high-pressure side of the refrigerant medium circuit, for example, viewed in the flow direction of the refrigerant through the refrigerant medium circuit, downstream of a refrigerant medium cooler, in particular designed as a condenser. However, even if the refrigerant medium collector is arranged on the low-pressure side, i.e., for example, between at least one evaporator and the suction side of the compressor, a reliable estimation of the refrigerant medium quantity can be made based on a comparison of at least one model parameter with at least one measured value.

[0027] This is advantageous because the refrigerant quantity present in the refrigerant circuit can be determined or determined very easily by generating at least one model which outputs at least one model parameter for a predefined or fixed refrigerant quantity.

[0028] Since the model only uses the at least one input parameter to output or predict at least one model parameter, the model also advantageously does not need to know the structural data of the refrigerant circuit. Instead, the specific structure of the corresponding refrigerant circuit is preferably incorporated into the model when the model is learned or trained by the input parameters present at the corresponding refrigerant quantity and the corresponding refrigerant circuit operating state.

[0029] Furthermore, the learning or training of the model can be automated very easily, so that this type of processing can also be realized easily and with little effort.

[0030] In particular, the model can be used very simply by a control device of a refrigerant circuit which is preferably configured to control components of the refrigerant circuit. Thus, the amount of refrigerant actually present in the refrigerant circuit can be estimated in real time and in continuous operation by means of the control device or controller.

[0031] By using the model and also by comparing the model parameters with the measured values, a prediction of at least one model parameter in the form of the refrigerant temperature at at least one predetermined location of the refrigerant circuit can be accomplished in a simple manner with a very low amount of calculations. This amount of calculations can be easily handled with the existing computing power resources in the control device or controller. This also facilitates the implementation of the method in a simple and low-cost manner.

[0032] Preferably, as at least one model parameter, the model outputs the temperature of the refrigerant at a location of the refrigerant circuit that is arranged upstream of the compressor and downstream of the at least one evaporator. This is because it has been shown that the total amount of refrigerant actually present in the refrigerant circuit can be inferred very well from the difference between the temperature predicted or output by the model and the actual or measured refrigerant temperature upstream of the compressor. Thus, by creating a temperature model for the refrigerant before or upstream of the compressor when a specific amount of refrigerant is present in the refrigerant circuit, a model parameter that can be used for comparison can be predicted or output by the model. And advantageously, the refrigerant temperature upstream of the compressor can be used to accurately determine the actual amount of refrigerant.

[0033] Additionally or alternatively, the model can output as at least one model parameter the refrigerant temperature at a location of the refrigerant circuit which is arranged downstream of the refrigerant cooler and upstream of the at least one expansion device. This is because the difference between the model-based temperature (i.e. the model parameter output by the model) and the refrigerant temperature measured downstream of the refrigerant cooler, which is in particular designed as a condenser, can be used to infer very well the total amount of refrigerant present in the refrigerant circuit, in particular based on the mutual relationship.

[0034] In particular, if not only the refrigerant temperature upstream of the compressor but also the refrigerant temperature downstream of the refrigerant cooler is used as corresponding model parameters for comparison with corresponding measured values, a particularly reliable estimate of the amount of refrigerant actually present in the refrigerant circuit can be achieved by comparing such model parameters with the measured values.

[0035] Preferably, the model outputs at least one model parameter based on a preset amount of refrigerant. Accordingly, the model can output at least one model parameter based on a fixed or preset amount of refrigerant, or predict the refrigerant temperature present at at least one predetermined location of the refrigerant circuit. By taking the preset amount of refrigerant as a basis when predicting at least one model parameter, the amount of refrigerant actually present in the refrigerant circuit can be inferred very well and simply based on the difference between the model parameter and the measured value.

[0036] Preferably, if the measured value of the temperature is greater than a model parameter output by the model as a function of the preset amount of refrigerant, it is inferred that the amount of refrigerant actually present in the refrigerant circuit is less than the amount of refrigerant on which the model is based. Since in this case it can be assumed that there is a lack of cooling medium, which is manifested by a measured temperature that is higher than the predicted temperature in the form of the model parameter. Thus, reliable conclusions can be drawn about the amount of refrigerant actually present in the refrigerant circuit.

[0037] The basis for this is the knowledge that, for example, due to refrigerant losses in the refrigerant circuit, i.e., when the amount of refrigerant present in the refrigerant circuit is less than the amount of refrigerant provided for the refrigerant circuit, the compressor delivers a smaller amount of refrigerant to the at least one evaporator, which in turn results in the refrigeration capacity provided by the at least one evaporator, in particular the outlet temperature of the air guided via the evaporator or the ability of the evaporator to absorb heat from the cooling medium flow being lower than the refrigeration capacity that should be provided or required.

[0038] If, in contrast, a predetermined amount of refrigerant is present in the refrigerant circuit, this reduction in the refrigeration capacity of the at least one evaporator can be counteracted by delivering more refrigerant to the at least one evaporator via the compressor. However, if too little refrigerant is present in the refrigerant circuit overall, even a transient increase in the delivery capacity of the compressor may not lead to a corresponding increase in the refrigeration capacity of the evaporator.

[0039] Preferably, if the measured value of the temperature is below a model parameter output by the model as a function of a predefined refrigerant quantity, it is concluded that the refrigerant quantity actually present in the refrigerant circuit is greater than the refrigerant quantity on which the model is based.

[0040] That is, if there is more refrigerant actually present in the refrigerant circuit than the refrigerant originally predicted by the model, i.e., based on or preset for the output of the model parameter by the model, then a state may occur in which the measured value of the temperature is lower than the model parameter output by the model according to the preset refrigerant amount. Therefore, the corresponding inference reflects the situation of the refrigerant amount actually present in the refrigerant circuit very well.

[0041] Preferably, the at least one input parameter is fed to a plurality of models, wherein the respective models output model parameters as a function of the respective differently predefined refrigerant quantities. In this case, the refrigerant quantity actually present in the refrigerant circuit is estimated taking into account the model parameters output by the respective models. By using a plurality of models for outputting a predicted refrigerant temperature, i.e., the corresponding model parameters, when a plurality of refrigerant quantities are predefined, a particularly accurate conclusion about the refrigerant quantity actually present can be drawn.

[0042] Because, taking into account the degree of deviation of the model parameters and the measured values ​​from one another, it is possible to infer the difference between the amount of refrigerant on which the model is based, which is preset or fixed, and the amount of refrigerant actually present in the refrigerant circuit. In this way, not only a qualitative conclusion is achieved about whether the amount of refrigerant actually present in the refrigerant circuit is more or less than the amount of refrigerant on which the model is based. Furthermore, a quantitative conclusion can be drawn about the amount of refrigerant actually present in the refrigerant circuit. This is advantageous when taking measures to solve the problem of a possible insufficient amount of refrigerant or a loss of refrigerant or an overfilling of the refrigerant circuit with refrigerant.

[0043] In particular, the amount of refrigerant actually present in the refrigerant circuit can be classified by referring to the preset amount of refrigerant for which the corresponding model has been trained. For example, the classification result can indicate whether the amount of refrigerant actually present in the refrigerant circuit is normal or abnormal.

[0044] Furthermore, particularly reliable quantitative conclusions can be drawn about the actual refrigerant quantity. For example, a corresponding quantification can be achieved by interpolating the classification result using a predefined refrigerant quantity.

[0045] Preferably, in order to estimate the amount of refrigerant actually present in the refrigerant circuit, the difference between the model parameter output by the corresponding model and the measured value is compared with the table value. By providing a corresponding logic table, the amount of refrigerant actually present in the refrigerant circuit can be determined very accurately and reliably. In addition, such a logic table in which the table values ​​are stored can be implemented very simply in a control device that outputs at least one model parameter when using at least one model. In order to determine the amount of refrigerant actually present, in particular a linear interpolation or also a nonlinear interpolation can be performed based on the table value.

[0046] Preferably, a neural network is used to output at least one model parameter, before which at least one input parameter for a fixed or predefined amount of refrigerant is supplied as training data to the neural network. Such a neural network can be implemented very easily and with little effort in a control device and requires only a very low computing power. Thus, even with relatively limited resources in terms of computing power, as can be the case, for example, in a control device for controlling components of a refrigerant circuit, the amount of refrigerant actually present in the refrigerant circuit can be calculated very easily.

[0047] As training data, simulation data and / or data obtained on a test bench and / or data obtained during operation of a motor vehicle equipped with the refrigerant circuit can be used. In addition to such data-based models, it can also or alternatively be provided that the model has an analytical nature, i.e., other data are derived from existing data, wherein the existing data and / or the derived data can be fed to the model as training data.

[0048] In particular, a particularly high prediction quality can be achieved by using a neural network to output at least one model parameter.

[0049] By feeding training data to the neural network, the model can be trained, wherein the trained model can then be used to calculate or determine an estimate of the amount of refrigerant medium actually present in the refrigerant medium circuit.

[0050] The calculation algorithm of the neural network which can be used to output at least one model parameter can be implemented very simply as software in the control device. The artificial network is learned or trained during the software development of a device for a corresponding refrigerant circuit, in particular one which can be used in a motor vehicle.

[0051] Preferably, a lookup table is used to provide at least one activation function of at least one neuron of the neural network. The use of at least one lookup table to provide at least one activation function, in particular a step function, of at least one neuron of the neural network ensures that the number of calculations required by the control device using the neural network for outputting at least one model parameter is particularly low. This is advantageous.

[0052] In order to output at least one model parameter, in addition to or as an alternative to the neural network, for example, at least one polynomial function of the system behavior to be predicted can be obtained by regression analysis of the data detected by simulation or in operation of the refrigerant circuit, for example using a polynomial model. At least one model parameter can also be output by means of such a polynomial function for subsequently estimating the actual refrigerant quantity based on a comparison of the model parameter with the measured value.

[0053] Preferably, the model is fed with the ambient temperature and / or the instantaneous delivery power (such as, for example, the rotational speed of the compressor) and / or the refrigerant pressure present on the suction side of the compressor as at least one input parameter relevant to the operation of the refrigerant circuit. It has been shown that based on such input parameters, the refrigerant temperature at a predetermined location can be output very simply and reliably by the model as a predicted model parameter. In addition, such input parameters are always available in a low-cost manner during the operation of the refrigerant circuit, so that their use by the model is particularly advantageous.

[0054] Preferably, the amount of refrigerant is estimated by a control device of the refrigerant circuit based on a comparison of model parameters with measured values. Preferably, the control device controls at least one component of the refrigerant circuit during operation of the refrigerant circuit. The method can be implemented with particularly low effort if the control device, which is to be provided for the actuation of at least one component of the refrigerant circuit, is also used for estimating the amount of refrigerant.

[0055] In particular, the control device can actuate, as at least one component, a compressor and / or at least one expansion device and / or at least one fan of the refrigerant circuit or similar units that can be used in the operation of the refrigerant circuit.

[0056] The motor vehicle according to the invention comprises a refrigerant circuit. A compressed refrigerant can be delivered to at least one refrigerant cooler of the refrigerant circuit by means of a compressor of the refrigerant circuit. The refrigerant from the refrigerant cooler can be expanded by means of at least one expansion mechanism of the refrigerant circuit. The expanded refrigerant can be delivered to at least one evaporator of the refrigerant circuit. The refrigerant from at least one evaporator can be delivered to the suction side of the compressor. At least one input parameter related to the operation of the refrigerant circuit can be delivered to a model, wherein the model can be used by a control device of the refrigerant circuit. The model is designed to output at least one model parameter, wherein the at least one model parameter can be used to determine the amount of refrigerant present in the refrigerant circuit. The model is designed to output the temperature of the refrigerant at at least one predetermined location of the refrigerant circuit as at least one model parameter. The control device is designed to compare the model parameter with a measured value of the refrigerant temperature detected at a predetermined location of the refrigerant circuit, and to estimate the amount of refrigerant present in the refrigerant circuit based on the comparison of the model parameter with the measured value.

[0057] By means of the control device, the amount of refrigerant actually present in the refrigerant circuit can be estimated with particularly high accuracy. Accordingly, a motor vehicle having the refrigerant circuit and the control device is designed to carry out the method according to the invention.

[0058] The advantages and preferred embodiments described for the method according to the invention also apply to the motor vehicle according to the invention, and vice versa.

[0059] Accordingly, the present invention also includes improvements of the motor vehicle according to the present invention, which have the features already described in conjunction with the improvements of the method according to the present invention. For this reason, the corresponding improvements of the motor vehicle according to the present invention will not be described in detail here.

[0060] The motor vehicle according to the invention is preferably designed as a car, in particular a passenger car or a commercial vehicle, or as a bus or a motorcycle.

[0061] For use cases or application scenarios that can be obtained in the method and are not explicitly described here, it can be provided that an error message is output according to the method.

[0062] The invention also includes a control device for a motor vehicle. The control device can have a data processor device or a processor device, which is configured to execute an embodiment of the method according to the invention. For this purpose, the processor device can have at least one microprocessor and / or at least one microcontroller and / or at least one FPGA (Field Programmable Gate Array) and / or at least one DSP (Digital Signal Processor). In addition, the processor device can have a program code, which is configured to execute an embodiment of the method according to the invention when executed by the processor device. The program code can be stored in a data memory of the processor device.

[0063] As another solution, the present invention also includes a computer-readable storage medium, which includes a command, which, when executed by a computer or a computer group, causes the execution of an embodiment of the method according to the present invention. For example, the storage medium can be at least partially designed as a non-volatile data storage (such as flash memory and / or SSD-solidstate drive, solid state hard disk) and / or at least partially designed as a volatile data storage (such as RAM-randomaccess memory, random access memory). A processor circuit with at least one microprocessor can be provided by a computer or a computer cluster. The command can be provided as a binary code or as an assembler and / or as a source code of a programming language (such as C language).

[0064] The present invention also includes combinations of features from the embodiments described. Therefore, the present invention also includes implementations which each have a combination of features from a plurality of the embodiments described, as long as these embodiments are not described as mutually exclusive. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] The following describes an embodiment of the present invention.

[0066] Figure 1 A motor vehicle with a refrigerant medium circuit is shown highly schematically; and

[0067] Figure 2 The use of a neural network to determine the amount of refrigerant medium actually present in a refrigerant medium circuit is schematically illustrated. DETAILED DESCRIPTION

[0068] The embodiments described below are preferred embodiments of the present invention. In the embodiments, the described parts of the embodiments are the features of the present invention that can be regarded as independent of each other, and these features also improve the present invention independently of each other. Therefore, the present disclosure should also include feature combinations that are different from the feature combinations of the illustrated embodiments. In addition, the embodiments can also be supplemented by other features in the features already described of the present invention.

[0069] In the figures, the same reference numerals respectively denote elements with the same function.

[0070] exist Figure 1 FIG. 1 schematically shows a motor vehicle 10 having a refrigerant circuit 12, wherein the refrigerant circuit Figure 1 1 and 2. The refrigerant circuit 12 is also only shown highly schematically. The refrigerant circuit 12 comprises a compressor 14, which compresses the sucked-in refrigerant and delivers the compressed refrigerant to a refrigerant cooler 16, which can be designed as a condenser, for example. When the refrigerant circuit 12 is in operation, the refrigerant is cooled in the refrigerant cooler 16, for example by means of air from the surroundings 18 of the motor vehicle 10.

[0071] A refrigerant collector 20 is preferably arranged in the refrigerant circuit 12, whereby the refrigerant collector is connected to the refrigerant circuit 12 on the high pressure side, i.e., arranged upstream of an expansion device 22 of the refrigerant circuit 12. For example, the refrigerant collector 20 can be integrated into the refrigerant cooler 16.

[0072] The refrigerant expanded by means of the expansion device 22 is supplied to the evaporator 24 of the refrigerant circuit 12. The air flowing through the evaporator 24 can be cooled by the evaporated refrigerant absorbing heat from the air. The cooled air can then be introduced into the passenger compartment 26 of the motor vehicle 10, which is only schematically shown here.

[0073] In addition to the schematically shown evaporator 24, the refrigerant circuit 12 can additionally or alternatively include an evaporator designed as a so-called refrigerator and having a further (not shown) expansion mechanism, by means of which the cooling medium flow can be cooled due to the evaporation of the refrigerant. The cooling medium cooled by means of the refrigerator can be used in particular for cooling a battery (not shown) of the motor vehicle 10.

[0074] During operation of the refrigerant circuit 12 , the refrigerant is supplied from the at least one evaporator 24 to the suction side 28 of the compressor 14 again.

[0075] exist Figure 1, a control device 30 is schematically shown which is configured to control components of the refrigerant medium circuit 12. For example, the rotational speed of the compressor 14 and thus the instantaneous delivery power of the compressor 14 can be preset by means of the control device 30. It is also preferably possible to control at least one expansion mechanism 22 of the refrigerant medium circuit 12 by means of the control device 30, i.e., for example, to cause the expansion mechanism 22 to open and close or to adjust the opening width of the expansion mechanism 22. Additionally or alternatively, it can be provided that the control device 30 controls a ventilator (not shown) assigned to the condenser 16, for example, by presetting the rotational speed of the ventilator. For example, a corresponding module 58, in particular a software module for controlling such components of the refrigerant medium circuit 12, is implemented in the control device 30. Figure 2 It is schematically shown in FIG.

[0076] Particularly in the refrigerant circuit 12 shown here by way of example with a refrigerant collector 20, it has proven difficult to determine the refrigerant quantity present in the refrigerant circuit 12. Here, for this purpose, the refrigerant quantity is estimated indirectly and based on a model.

[0077] For this purpose, input parameters 34 are supplied to at least one model 32, which is implemented in this case in a control device 30 (see Figure 2 ). The model 32 then outputs at least one model parameter 36. The model parameter 36 is compared with the measured value 38. Based on this comparison 40, the method implemented in the control device 30 or in such a controller estimates the amount of refrigerant 42 actually present in the refrigerant circuit 12 (see Figure 2 ).

[0078] Thus, in the indirect processing used here, the model 32 does not directly output the actual refrigerant amount 42, but rather outputs a model parameter 36, which is a parameter that can be measured when the refrigerant circuit 12 is operating normally. At least one model 32 is created for the at least one model parameter 36. The difference between the measured parameter, i.e., the measured value 38, which occurs when the refrigerant circuit 12 is operating normally, and the model parameter 36 predicted by the model 32 is used to estimate the refrigerant amount 42 actually present in the refrigerant circuit 12.

[0079] It has proven to be advantageous if, for example, the temperature of the refrigerant downstream of the compressor 14 can be output as at least one model parameter 36. A corresponding measured value 38 of the temperature present at such a predetermined point in the refrigerant circuit 12 can be measured, for example, by means of a first temperature sensor 44 of the refrigerant circuit 12, which is located at a predetermined position of the refrigerant circuit 12. Figure 1If the model parameter 36 is now compared with a measured value 38 , for example in the form of a temperature detected upstream of the compressor 14 by means of a temperature sensor 44 , then an estimate 42 of the refrigerant medium quantity can be made based on this comparison 40 .

[0080] Additionally or alternatively, at least one model 32 can output, as at least one model parameter 36, a temperature of the refrigerant downstream of the refrigerant cooler 16 and thus upstream of the expansion device 22. The refrigerant temperature actually present at a predetermined location of the refrigerant circuit 12 can be measured by means of a further temperature sensor of the refrigerant circuit 12 or a second temperature sensor 46, which is also located at a predetermined location of the refrigerant circuit 12. Figure 1 The further temperature sensor 46 then provides a measured value 38 which is compared 40 with the model parameter 36 (see Figure 2 ).

[0081] It has proven to be advantageous that in particular not only the difference between the model parameter 36 and the measured value 38 of the refrigerant temperature before the compressor 14, but also the temperature difference between the predicted model parameter 36 and the actual, measured value 38 of the refrigerant temperature after the refrigerant cooler 16, which is designed in particular as a condenser, can be used very well to infer the total amount of refrigerant 42 actually present in the refrigerant circuit 12. This is because these differences are related to the difference between the refrigerant amount on which the model 32 is based and the refrigerant amount 42 actually present in the refrigerant circuit 12. Advantageously, the temperature is predicted based on the input parameter 34 during the continuous operation of the refrigerant circuit 12, i.e. the model parameter 36 is output.

[0082] In addition to the use of model parameters 36 and measured values ​​38 of the temperature, it is additionally or alternatively conceivable to use at least one further value, which can be predicted on the one hand by means of at least one model 32 and can be detected on the other hand by measurement technology, in order to estimate the amount of refrigerant 42 actually present in the refrigerant circuit 12. For example, the at least one further value can be the pressure present at at least one predetermined point of the refrigerant circuit 12.

[0083] The at least one model 32 can be implemented in the control device or in the control device 30 for outputting at least one model parameter 36. In this case, the at least one model 32 and the algorithm for the comparison 40 are implemented in the control device 30 (see Figure 2 ).

[0084] The result of comparison 40, i.e., estimated refrigerant quantity 42 present in refrigerant circuit 12, can be used by control device 30 to trigger at least one response, in particular in the event of a lack of refrigerant. For example, the at least one response can consist in outputting a warning to a user of motor vehicle 10 and / or switching off compressor 14.

[0085] Preferably, the corresponding model 32 outputs at least one model parameter 36 according to the corresponding preset refrigerant medium amount. In other words, preferably, each model in the model 32 is based on a corresponding different refrigerant medium amount, and outputs a corresponding model parameter 36 based on the based or preset refrigerant medium amount.

[0086] For example, the first model 32 can output a model parameter 36 based on a fixed or preset refrigerant amount, which is compared with a measured value 38 of the temperature upstream of the compressor 14. Another model or second model 32 can output a predicted temperature downstream of the refrigerant cooler 16 as a model parameter 36, wherein the second model 32 is also based on a preset refrigerant amount (for training the model 32).

[0087] By using a corresponding temperature model for outputting corresponding model parameters 36, a particularly high degree of reliability can be achieved when estimating the actual refrigerant quantity 42. Accordingly, the temperature model can output the refrigerant temperature before the compressor 14 or after the refrigerant cooler 16 or condenser as a corresponding model parameter 36 and is created in each case under a constant or fixedly preset refrigerant quantity and the operating characteristics of the refrigerant circuit 12 resulting therefrom.

[0088] In order to output the at least one model parameter 36, a neural network 48 can be used in particular (see Figure 2 ), the neural network can be implemented in the control device 30, and according to Figure 2 , the neural network can have a plurality of neurons 50. For example, the ambient temperature (i.e. the temperature in the surroundings 18 of the motor vehicle 10), the instantaneous delivery capacity of the compressor 14 and the refrigerant pressure present at the suction side 28 of the compressor 14 can be supplied to the neural network 48 as input variables 34.

[0089] For example, the control device 30 can use the rotational speed of the compressor 14 as one of the input parameters 34. The ambient temperature can be determined by the flow of air in the refrigerant circuit 12. Figure 1 Detection is performed by a temperature sensor 52 schematically shown in the figure, which can detect, for example, the temperature of the supply air or the ambient air that is fed to the refrigerant cooler 16. In order to detect the suction pressure, that is, the refrigerant pressure existing at the suction side 28 of the compressor 14 in the refrigerant circuit 12, a temperature sensor 52 can be used. Figure 1 A pressure sensor 54 of the refrigerant circuit 12 is schematically shown in FIG. 5 , which is arranged upstream of the compressor 14 .

[0090] Here, the model 32 predicts at least one model parameter 36 from such input parameters 34 - in particular using a neural network 48 - for a preset or fixed refrigerant quantity on which the model 32 is based, i.e., for example, the refrigerant temperature before the compressor 14 and / or the refrigerant temperature after the refrigerant cooler 16 .

[0091] For this purpose, the at least one model 32 has already been trained in a previously performed training phase using training data 56. Accordingly, the input parameters 34 and a fixedly predefined amount of refrigerant can be supplied to the respective model 32 as training data 56. Subsequently, the at least one trained model 32 outputs at least one model parameter 36, which is compared with the respective measured value 38.

[0092] It may be the case that the amount of refrigerant 42 actually present in the refrigerant circuit 12 is greater than the predefined amount of refrigerant on which the output of the model parameter 36 is based. If the amount of refrigerant 42 actually present in the refrigerant circuit 12 is greater than the amount of refrigerant assumed by the model 32 for which the model 32 has been trained, then the actually measured refrigerant temperature, for example the refrigerant temperature upstream of the compressor 14, and thus the measured value 38, is also lower than the temperature predicted by the model in the form of the model parameter 36.

[0093] If, on the other hand, the amount of refrigerant 42 actually present in the refrigerant circuit 12 is less than the amount of refrigerant assumed by the model 32 and for which the model was trained, the actually measured temperature, i.e. the measured value 38, is greater than the model parameter 36 predicted by the model 32. These dependencies apply not only to the refrigerant temperature detected by means of a first temperature sensor 44 of the refrigerant circuit 12 upstream of the compressor, but also to the refrigerant temperature detected downstream of the refrigerant cooler 16 by means of a second temperature sensor 46 of the refrigerant circuit 12.

[0094] The model 32 for the refrigerant temperature after the refrigerant cooler 16 can be used alone or in combination with the model 32 for the refrigerant temperature before the compressor 14 to determine the refrigerant quantity 42 actually present in the refrigerant circuit 12. Figure 2 Only one of these models 32 is shown.

[0095] The greater the difference between the model prediction (i.e. the at least one model parameter 36) and the measured temperature (i.e. the measured value 38), the greater the difference between the amount of refrigerant assumed by the at least one model 32 or on which the at least one model 32 is based and the amount of refrigerant 42 actually present in the refrigerant circuit 12. Accordingly, it is possible to draw not only qualitative conclusions about the amount of refrigerant 42 actually present in the refrigerant circuit 12, but also quantitative conclusions.

[0096] Particularly accurate conclusions can be drawn if a plurality of models 32 are used to output individual model parameters 36 in the form of predicted refrigerant temperatures, with different or mutually different refrigerant quantities assumed, predefined or fixed when training the respective models 32. This makes it possible to classify the actual refrigerant quantities 42 according to the refrigerant quantities for which the used models 32 have been trained.

[0097] according to Figure 2 , an estimate for the current or actual refrigerant quantity 42 or filling level is calculated with the aid of a neural network 48. The neural network 48 can be trained for each type of motor vehicle 10 or for each type of refrigerant circuit 12 installed in a motor vehicle 10, for example during software development of other devices of the motor vehicle 10.

[0098] As schematically shown in Figure 2 As shown in , a corresponding calculation algorithm of the neural network 48 can be implemented in the software of the control unit or control device 30 .

[0099] Furthermore, the step functions or activation functions of the individual neurons 50 present in the neural network 48 can be stored in a lookup table for calculation in the control device 30. In this way, even with relatively limited resources in terms of computing power, such as may be the case in the control device 30, the calculation or estimation of the actual refrigerant quantity 42 can be realized without any problems.

[0100] Overall, the example shows that an improved estimation of the refrigerant quantity 42 in the refrigerant circuit 12 or in a corresponding refrigerant system can be made based on the model.

Claims

1. A method for determining a refrigerant quantity (42) in a refrigerant circuit (12) of a motor vehicle (10), in which method a compressor (14) of the refrigerant circuit (12) delivers compressed refrigerant to at least one refrigerant cooler (16) of the refrigerant circuit (12), wherein: The refrigerant from the refrigerant cooler (16) is expanded by means of at least one expansion device (22) of the refrigerant circuit (12), wherein the expanded refrigerant is supplied to at least one evaporator (24) of the refrigerant circuit (12), wherein the refrigerant from the at least one evaporator (24) is supplied to the suction side (28) of the compressor (14), wherein at least one input parameter (34) related to the operation of the refrigerant circuit (12) is supplied to a model (32) which outputs at least one model parameter (36), wherein the at least one model parameter (36) is used to determine the amount of refrigerant (42), characterized in that The model (32) outputs, as the at least one model parameter (36), a refrigerant temperature at a predetermined location of the refrigerant circuit (12), wherein the model parameter (32) is compared with a measured value (38) of the temperature of the refrigerant detected at the predetermined location of the refrigerant circuit (12), wherein the refrigerant amount (42) is estimated based on the comparison (40) of the model parameter (36) and the measured value (38).

2. The method according to claim 1, characterized in that As the at least one model parameter (36), the model (32) outputs the refrigerant temperature at a location of the refrigerant circuit (12) arranged upstream of the compressor (14) and downstream of the at least one evaporator (24) and / or the refrigerant temperature at a location of the refrigerant circuit (12) arranged downstream of the refrigerant cooler (16) and upstream of the at least one expansion mechanism (22).

3. The method according to any one of the preceding claims, characterized in that The model (32) outputs the at least one model parameter (36) according to a preset amount of refrigerant.

4. The method according to claim 3, It is characterized in that If the measured value (38) of the temperature is greater than a model parameter (36) output by the model (32) based on a preset refrigerant quantity, it is inferred that the refrigerant quantity (42) actually present in the refrigerant circuit (12) is less than the refrigerant quantity on which the model (32) is based, wherein: If the measured value (38) of the temperature is below a model parameter (36) output by the model (32) based on a preset refrigerant quantity, it is inferred that the refrigerant quantity (42) actually present in the refrigerant circuit (12) is greater than the refrigerant quantity on which the model (32) is based.

5. The method according to claim 3 or 4, characterized in that: The at least one input parameter (34) is supplied to a plurality of models (32), wherein each model (32) outputs a model parameter (36) as a function of differently predetermined refrigerant quantities, wherein the refrigerant quantity (42) actually present in the refrigerant circuit (12) is estimated taking into account the model parameter (36) output by the corresponding model (32).

6. The method according to any one of claims 3 to 5, characterized in that A neural network (48) is used to output the at least one model parameter (36), and before the at least one model parameter (36) is output, at least one input parameter (34) at a predetermined amount of refrigerant is supplied to the neural network as training data (56).

7. The method according to claim 5 or 6, characterized in that: In order to estimate the amount of refrigerant (42) actually present in the refrigerant circuit (12), the difference between a model parameter (36) output by a corresponding model (32) and a measured value (38) is compared with a table value and / or a lookup table is used to provide at least one activation function of at least one neuron (50) of a neural network (48).

8. The method according to any one of the preceding claims, characterized in that As at least one input variable (34) relevant to the operation of the refrigerant circuit (12), the model (32) is fed with the ambient temperature and / or the instantaneous delivery capacity of the compressor (14) and / or the refrigerant pressure prevailing on the suction side (28) of the compressor (14).

9. The method according to any one of the preceding claims, characterized in that A control device (30) of a refrigerant circuit (12) estimates a refrigerant quantity (42) based on a comparison (40) of a model parameter (36) with a measured value (38), the control device controlling at least one component of the refrigerant circuit (12) when the refrigerant circuit (12) is in operation.

10. A motor vehicle having a refrigerant medium circuit (12), wherein: A refrigerant can be compressed and delivered to at least one refrigerant cooler (16) of the refrigerant circuit (12) by means of a compressor (14) of the refrigerant circuit (12), wherein the refrigerant from the refrigerant cooler (16) can be expanded by means of at least one expansion device (22) of the refrigerant circuit (12), wherein the expanded refrigerant can be delivered to at least one evaporator (24) of the refrigerant circuit (12), and the refrigerant from the at least one evaporator (24) can be delivered to a suction side (28) of the compressor (14), wherein at least one input parameter (34) related to the operation of the refrigerant circuit (12) can be delivered to a model (32) that can be used by a control device (30) of the refrigerant circuit (12), wherein the model (32) is designed to output at least one model parameter (36), wherein the at least one model parameter (36) can be used to determine the amount (42) of refrigerant present in the refrigerant circuit (12), It is characterized in that The model (32) is designed to output, as the at least one model parameter (36), a refrigerant temperature at a predetermined location of the refrigerant circuit (12), wherein the control device (30) is designed to compare the model parameter (36) with a measured value (38) of the refrigerant temperature detected at the predetermined location of the refrigerant circuit (12), and to estimate the amount of refrigerant (42) present in the refrigerant circuit (12) based on the comparison (40) of the model parameter (36) with the measured value (38).

Images